JP4053460B2 - Image processing apparatus, image forming apparatus, image processing method, image processing program, and recording medium - Google Patents

Description

【００６０】
次に、ステップＳ５では、最初に、被処理３×３ブロック群の分散平均値ＡＬＬＶＡＲが算出される。３×３ブロック群の分散平均値ＡＬＬＶＡＲは、該３×３ブロック群内の全ブロックの分散Ｖａｒ(１)〜Ｖａｒ(９)の平均値である。分散平均値ＡＬＬＶＡＲは、次式に示すように、３×３ブロック群内の全ブロックの分散Ｖａｒ(１)〜Ｖａｒ(９)の和ＡＬＬＳＵＭを３×３ブロック群内の全ブロックの数で除算したものである。
【００６１】
【数２】

【００６２】
次いで、ステップＳ５では、被処理３×３ブロック群の分散平均値ＡＬＬＶａｒと、閾値格納部２３に予め格納されている第１ノイズ判定閾値ＴＨ１とが比較される。分散平均値ＡＬＬＶａｒが第１ノイズ判定閾値ＴＨ１以下であれば、画像内においてノイズが生じる可能性のある平坦部に被処理３×３ブロック群の注目ブロックＢ５が位置すると判定して、ステップＳ５からステップＳ６に進む。
【００６３】
ステップＳ６では、被処理３×３ブロック群の全ブロックの分散Ｖａｒ(１)〜Ｖａｒ(９)と、閾値格納部２３に予め格納されている第２ノイズ判定閾値ＴＨ２とが比較される。被処理３×３ブロック群の全ブロックの分散Ｖａｒ(１)〜Ｖａｒ(９)が第２ノイズ判定閾値ＴＨ２よりも大きければ、ステップＳ６からステップＳ１０に進む。被処理３×３ブロック群の全ブロックの分散Ｖａｒ(１)〜Ｖａｒ(９)が第２ノイズ判定閾値ＴＨ２以下であれば、ステップＳ６からステップＳ１１に進む。
【００６４】
ステップＳ６における比較処理は、具体的には、被処理３×３ブロック群の各ブロックの分散Ｖａｒ(１)〜Ｖａｒ(９)を第２ノイズ判定閾値ＴＨ２とそれぞれ比較することによって行われる。すなわち、被処理３×３ブロック群の全ブロックの分散Ｖａｒ(１)〜Ｖａｒ(９)が第２ノイズ判定閾値ＴＨ２より大きいとは、ブロックの分散Ｖａｒ(１)〜Ｖａｒ(９)の全てが第２ノイズ判定閾値ＴＨ２より大きいことを意味する。このように、ステップＳ６では、全ブロックの分散Ｖａｒ(１)〜Ｖａｒ(９)と第２ノイズ判定閾値ＴＨ２とを比較して、注目ブロックに生じているブロックノイズのレベルを判定している。すなわち、全ブロックの分散Ｖａｒ(１)〜Ｖａｒ(９)が大きいということは、上記３×３ブロック群に生じているブロックノイズのレベルが高いことを意味する。
【００６５】
全ブロックの分散が第２ノイズ判定閾値ＴＨ２よりも大きい場合、ステップＳ１０において、第１ブロックノイズ除去フィルタＦ１を用いる空間フィルタ処理を注目ブロックを構成する各画素に対して施すことによって、平滑化が行われる。第１ブロックノイズ除去フィルタＦ１は、たとえば図６（Ａ）に示すような、比較的強い効果のあるブロックノイズ除去用のフィルタである。全ブロックの分散が第２ノイズ判定閾値ＴＨ２以下である場合、ステップＳ１１において、第２ブロックノイズ除去フィルタＦ２を用いる空間フィルタ処理を注目ブロックを構成する各画素に対して施すことによって、平滑化が行われる。第２ブロックノイズ除去フィルタＦ２は、たとえば図６（Ｂ）に示すような、比較的弱い効果のあるブロックノイズ除去用のフィルタであり、少なくとも、第１ブロックノイズ除去フィルタＦ１よりも平滑化効果が弱い。
【００６６】
前述のステップＳ５において、被処理３×３ブロック群の分散平均値ＡＬＬＶａｒが第１ノイズ判定閾値ＴＨ１より大きい場合、画像内においてリンギングノイズが生じる可能性のあるエッジ部に被処理３×３ブロック群の注目ブロックＢ５が位置するか否かを判定するために、ステップＳ５からステップＳ７に進む。リンギングノイズの有無は、ステップＳ７およびステップＳ８で２段階に判定される。
【００６７】
ステップＳ７では、被処理３×３ブロック群の全ブロックのうち、注目ブロックＢ５を除く８つの周辺ブロックＢ１〜Ｂ４,Ｂ６〜Ｂ９の分散Ｖａｒ(１)〜Ｖａｒ(４),Ｖａｒ(６)〜Ｖａｒ(９)のうちの少なくとも１つが、第１ノイズ判定閾値ＴＨ１未満であるか否かが判定される。さらに、ステップＳ７では、注目ブロックの分散Ｖａｒ(５)が第１ノイズ判定閾値ＴＨ１以上であるか否かが判定される。これらの判定結果に基づき、全周辺ブロックＢ１〜Ｂ４,Ｂ６〜Ｂ９の分散Ｖａｒ(１)〜Ｖａｒ(４),Ｖａｒ(６)〜Ｖａｒ(９)のうちの少なくとも１つが第１ノイズ判定閾値ＴＨ１未満であり、かつ、注目ブロックの分散Ｖａｒ(５)が第１ノイズ判定閾値ＴＨ１以上である場合、被処理３×３ブロック群内の注目ブロックＢ５にリンギングノイズが生じていると判定され、ステップＳ７からステップＳ１２に進む。周辺ブロックＢ１〜Ｂ４,Ｂ６〜Ｂ９の分散Ｖａｒ(１)〜Ｖａｒ(４)，Ｖａｒ(６)〜Ｖａｒ(９)が全て第１ノイズ判定閾値ＴＨ１以上である場合、または注目ブロックの分散Ｖａｒ(５)が第１ノイズ判定閾値ＴＨ１未満である場合、すなわちステップＳ７の条件に該当しない場合、ステップＳ７からステップＳ８に進む。
【００６８】
ステップＳ８では、ステップＳ７の条件とは異なる条件に基づいて、リンギングノイズの有無が再度判定される。具体的には、被処理３×３ブロック群の全ブロックの分散Ｖａｒ(１)〜Ｖａｒ(９)と該被処理３×３ブロック群の分散値の分散ＶａｒＶａｒとを用いて次式３に示すリンギングノイズの判定値が算出され、算出された該判定値に基づいてリンギングノイズの有無が判定される。３×３ブロック群の分散値の分散ＶａｒＶａｒは、次式４によって算出される。
【００６９】
【数３】

【００７０】
ステップＳ８では、式３に基づいて算出されたリンギングノイズの判定値と、閾値格納部２３に格納されている第４ノイズ判定閾値ＴＨ４とが比較される。第４ノイズ判定閾値ＴＨ４は、たとえば、０である。判定値が第４ノイズ判定閾値ＴＨ４よりも大であれば、リンギングノイズが生じていると判定されて、ステップＳ８からステップＳ１２に進む。判定値が第４ノイズ判定閾値ＴＨ４よりも小であれば、リンギングノイズは生じていないと判定されて、ステップＳ８からステップＳ９に進む。
【００７１】
ステップＳ７またはステップＳ８からステップＳ１２に進んだ場合、被処理３×３ブロック群の注目ブロックＢ５にはリンギングノイズが生じている。このため、ステップＳ１２において、たとえば、注目ブロックＢ５を構成する各画素に対して図６（Ｃ）に示すようなリンギングノイズ除去用のフィルタＦ３を用いる空間フィルタ処理を施すことによって、平滑化が行われる。
【００７２】
ステップＳ９では、ブロックノイズおよびリンギングノイズの双方が生じていない注目ブロックＢ５に対する平滑化処理に係る判定が行われる。ステップＳ９では、被処理３×３ブロック群のブロック間の連続性に基づいて、平滑化処理の強度が調整される。具体的には、注目ブロックの分散Ｖａｒ(５)が、予め設定されている第３ノイズ判定閾値ＴＨ３以下である場合には、ブロック間の連続性が比較的保たれていると判断され、ステップＳ９からステップＳ１３に進む。ステップＳ１３では、注目ブロックＢ５を構成する各画素に対して第１平滑化フィルタＦ４を用いる空間フィルタ処理を施すことによって、平滑化が行われる。第１平滑化フィルタＦ４は、たとえば図６（Ｄ）に示すような、比較的弱い効果のある平滑化処理用のフィルタである。
【００７３】
また、注目ブロックの分散Ｖａｒ(５)が第３ノイズ判定閾値ＴＨ３よりも大きいと判定された場合、ブロック間の連続性が比較的失われていると判断され、ステップＳ９からステップＳ１４に進む。ステップＳ１４では、注目ブロックＢ５を構成する各画素に対して第２平滑化フィルタＦ５を用いる空間フィルタ処理を施すことによって、平滑化が行われる。第２平滑化フィルタＦ５は、たとえば図６（Ｅ）に示すような、比較的強い効果のある平滑化処理用のフィルタであり、少なくとも第１平滑化フィルタＦ４よりも効果が強い。
【００７４】
以上説明したように、画像処理装置３は、３×３ブロック群を単位としてブロックノイズおよびリンギングノイズのノイズ判定を行い、ノイズ判定の結果に基づいてノイズの除去処理を行っている。画像処理装置３は、ステップＳ１０〜Ｓ１４のノイズ除去処理を行った後、各ブロック内の画素に対して中間調処理を施す。すなわち、図４のステップＳ１５〜ステップＳ１８において、ノイズの種類およびノイズのレベルに基づいて誤差拡散処理が行われる。
【００７５】
表１は、図４のステップＳ１５〜ステップＳ１８で用いられる量子化閾値ＴＨａの具体的な例を示す。ステップＳ１５〜ステップＳ１８および表１の説明では、画素の画素値が２５６階調で表されると仮定する。たとえば、画素値が２５６階調で表される場合、低濃度部の画素とは、０以上２０以下の画素値を有する画素である。高濃度部の画素とは、２３６以上２５５以下の画素値を有する画素である。また中濃度部の画素とは、２１以上２３５以下の画素値を有する画素である。
【００７６】
【表１】

【００７７】
ステップＳ１０において第１ブロックノイズ除去フィルタＦ１を用いてブロックノイズが除去された注目ブロックＢ５は、ステップＳ１０からステップＳ１５に進んで処理される。ステップＳ１５では、ブロックノイズが隣接するブロック間の境界に生じることに着目し、注目ブロックＢ５内の境界２画素領域ＡＲ１内の画素に対して、下記の誤差拡散処理を行う。境界２画素領域ＡＲ１は、図７において右上がりの斜線を付して示す領域であり、注目ブロックＢ５の境界部分に位置する２画素分の領域である。
【００７８】
ステップＳ１５の具体的な誤差拡散処理は以下のとおりである。復元画像内において、ステップＳ１０において処理された注目ブロックＢ５は、隣接するブロックＢ１〜Ｂ４，Ｂ６〜Ｂ９との間に連続性を有するので、復元画像内の平坦な部分に相当する。このため、まず、誤差拡散処理によって頻繁に画素値が反転しないように、境界２画素領域ＡＲ１のうちの低濃度および高濃度の画素に対して、図３に示す濃度判定部３４において、以下のような画素値の飽和処理が行われる。具体的には、低濃度部の画素に対しては、画素値が０に飽和させられる。また、高濃度部の画素に対しては、画素値が２５５に飽和させられる。中間濃度部の画素に対しては、画素値の飽和処理は施されず、元の画素値が保たれる。また、濃度判定部３４において画素値を飽和させる代わりに、濃度判定部３４の濃度判定信号に基づいて誤差判定部１３で誤差の切捨てまたは丸込みが行われてもよい。濃度判定部３４における飽和処理後、図３で説明したように、注目ブロックＢ５の各画素の画素値に対して、誤差格納部３３に格納されている拡散誤差が加算される。
【００７９】
次いで、ステップＳ５〜ステップＳ９において行われたノイズ判定によって得られたノイズ判定信号と、濃度判定部３４において行われた濃度判定によって得られた濃度判定信号とに基づいて、量子化処理部１２が量子化処理を行う。具体的には、ステップＳ１０において処理された注目ブロックＢ５は、比較的大きいブロックノイズが発生しているブロックであるので、量子化閾値格納部３１に格納された量子化閾値ＴＨａのうち、表１の「誤差拡散（１）」に示す量子化閾値ＴＨａを用いて量子化が行われる。
【００８０】
ステップＳ１５の量子化処理では、濃度判定部３４によって被処理画素が低濃度部または高濃度部の画素であると判定された場合、表１に示すように、レベル１の量子化閾値ＴＨａとして６４を用い、レベル３の量子化閾値ＴＨａとして１９２を用いる。また、濃度判定部３４によって被処理画素が中間濃度部の画素であると判定された場合、表１に示すように、レベル１の量子化閾値ＴＨａとして１８０を用い、レベル３の量子化閾値ＴＨａとして７６を用いる。これによって、中間濃度部の画素では、誤差拡散処理によって画素値を反転させやすくすることが可能になる。このように、ステップＳ１５の量子化処理では、中間濃度部の画素に対しては、誤差拡散処理による信号値の反転を起こりやすくするために、量子化閾値ＴＨａを高濃度側で低く、かつ低濃度側で高く設定する。
【００８１】
次いで、上述したように、拡散係数設定部１４が拡散係数ＴＨｃを決定する。拡散係数ＴＨｃは、ノイズ判定信号に基づいて定められる。ブロックノイズが発生する可能性のあるブロックに対しては、図８（Ａ）に示すような第１誤差拡散範囲Ｄｉｆ１の拡散係数ＴＨｃが設定され、第１誤差拡散範囲Ｄｉｆ１を用いて誤差拡散処理が行われる。
【００８２】
なお、上述の例では、低濃度部または高濃度部の画素の画素値を０または２５５に飽和させる場合について説明したが、誤差判定部１３が量子化誤差を０に補正するようにしてもよい。すなわち、濃度判定部３４が入力された画素の濃度を判定し、得られた濃度判定結果に基づいて、低濃度部または高濃度部と判定された画素については、誤差判定部１３において設定される判定誤差を０に設定してもよい。
【００８３】
また、ステップＳ１５の量子化処理では、注目ブロックＢ５内の境界２画素領域ＡＲ１以外の領域の画素に対しては、濃度判定部３４における濃度判定を行わずに、誤差拡散処理が行われる。すなわち、注目ブロックＢ５内の境界２画素領域ＡＲ１外の画素の濃度に応じて、量子化閾値ＴＨａを変化させて量子化することなく、誤差拡散処理を行えばよい。
【００８４】
ステップＳ１１において第２ブロックノイズ除去フィルタＦ２を用いてブロックノイズが除去された注目ブロックＢ５は、ステップＳ１１からステップＳ１６に進んで処理される。ステップＳ１６では、ブロックノイズが隣接するブロック間の境界に残りやすいことに着目し、注目ブロックＢ５内の境界１画素領域ＡＲ２内の画素に対して、下記の誤差拡散処理を行う。境界１画素領域ＡＲ２は、図７において右下がりの斜線を付して示す領域であり、注目ブロックＢ５の境界部分に位置する１画素分の領域である。すなわち、ステップＳ１６の処理は、ステップＳ１５で処理対象となった境界２画素領域ＡＲ１のうち、注目ブロックの境界側に位置する１画素分の領域に対して行われる。
【００８５】
ステップＳ１６の具体的な誤差拡散処理は以下のとおりである。まず、図３に示す濃度判定部３４が、ステップＳ１５で行った処理と同様に、濃度判定を行う。すなわち、低濃度部または高濃度部の画素は飽和させられ、中間濃度部の画素に対する飽和処理は行われない。次いで、表１の「誤差拡散（２）」に示す量子化閾値ＴＨａを用いて、量子化処理部１２が量子化を行う。量子化処理部１２は、ステップＳ１５で行った処理と同様に、低濃度部または高濃度部と中間濃度部とで、量子化閾値ＴＨａを変化させて量子化を行う。続いて、拡散係数設定部１４が、ノイズ判定信号に基づいて、図８（Ａ）に示す第１誤差拡散範囲Ｄｉｆ１の拡散係数ＴＨｃを設定し、第１誤差拡散範囲Ｄｉｆ１を用いて誤差拡散処理を行う。なお、ステップＳ１６においても、ステップＳ１５で説明したように、濃度判定部３４が低濃度部または高濃度部の画素の画素値を飽和させるのではなく、誤差判定部１３が量子化誤差を０に設定するようにしてもよい。ステップＳ１６の量子化処理では、注目ブロックＢ５内の境界１画素領域ＡＲ２以外の領域の画素に対しては、濃度判定部３４の濃度判定を行わずに、誤差拡散処理が行われる。すなわち、注目ブロックＢ５内の境界１画素領域ＡＲ２外の画素の濃度に応じて、量子化閾値ＴＨａを変化させて量子化することなく、誤差拡散処理を行えばよい。
【００８６】
ステップＳ１２においてリンギングノイズ除去フィルタＦ３を用いてブロックノイズが除去された注目ブロックＢ５は、ステップＳ１２からステップＳ１７に進んで処理される。リンギングノイズは、ブロック内の位置には関係なく、画素濃度のコントラストの高い部分に特に発生しやすいので、ステップＳ１７では、誤差拡散処理による誤差の蓄積によって画素値の反転が頻繁に起こりにくくなるように、処理が行われる。ステップＳ１７では、ブロック内の全画素に対して処理が行われる。
【００８７】
ステップＳ１７の具体的な誤差拡散処理は以下のとおりである。まず、濃度判定部３４が、ステップＳ１５およびステップＳ１６で行ったように、低濃度部または高濃度部の画素を飽和させ、中間濃度部の画素に対しては飽和処理を行わない。次いで、表１の「誤差拡散（３）」に示す量子化閾値ＴＨａを用いて、量子化処理部１２が量子化を行う。ステップＳ１７では、ステップＳ１５およびステップＳ１６と異なり、低濃度部と高濃度部と中間濃度部とに関係なく、注目ブロックＢ５内の全ての画素に対して、相互に同じ量子化閾値ＴＨａを用いて量子化が行われる。続いて、拡散係数設定部１４が、ノイズ判定信号に基づいて、図８（Ｂ）に示す第２誤差拡散範囲Ｄｉｆ２の拡散係数ＴＨｃを設定し、第２誤差拡散範囲Ｄｉｆ２を用いて誤差拡散処理を行う。第１誤差拡散範囲Ｄｉｆ１のほうが第２誤差拡散範囲Ｄｉｆ２よりも、処理すべき注目画素を中心とした処理範囲が広い。なお、ステップＳ１５およびステップＳ１６において説明したように、ステップＳ１７でも、濃度判定部３４が低濃度部または高濃度部の画素の画素値を飽和させるのではなく、誤差判定部１３が量子化誤差を０に設定するようにしてもよい。
【００８８】
また、ブロックノイズおよびリンギングノイズのいずれも生じていないと判定され、ステップＳ１３およびステップＳ１４において、第１および第２平滑化フィルタＦ４，Ｆ５を用いて平滑化処理された注目ブロックＢ５は、ステップＳ１３およびステップＳ１４からステップＳ１８に進んで処理される。ステップＳ１８に進んだ注目ブロックＢ５は歪みの生じていない領域内に位置するので、濃度判定部３４の濃度判定を行わず、表１の「誤差拡散（４）」に示す量子化閾値ＴＨａを用いて、量子化処理部１２が量子化を行う。つまり、量子化処理部１２は、低濃度部と高濃度部と中間濃度部に関係なく、注目ブロックＢ５内の全ての画素に対して相互に同じ量子化閾値ＴＨａを用いて量子化を行う。続いて、拡散係数設定部１４が、ノイズ判定信号に基づいて、図８（Ｂ）に示す第２誤差拡散範囲Ｄｉｆ２の拡散係数ＴＨｃを設定し、第２誤差拡散範囲Ｄｉｆ２を用いて誤差拡散処理を行う。ステップＳ１８の処理は、ステップＳ１７と同様、注目ブロックＢ５内の全画素に対して行われる。
【００８９】
このように、図４のステップＳ１５〜ステップＳ１８では、ノイズ判定信号に基づいて、ノイズの種類およびノイズのレベルに応じた誤差拡散処理を行うとともに、濃度判定部３４で判定された濃度判定信号にも基づいて誤差拡散処理を行っている。これによって、ステップＳ４〜ステップＳ１４で行われる空間フィルタ処理部９のノイズ除去処理だけではなく、階調再現処理部８で行われる誤差拡散処理に際しても、量子化閾値ＴＨａおよび誤差判定閾値ＴＨｂあるいは拡散係数ＴＨｃを切換えるという簡易な手法を用いて、ノイズの種類およびノイズのレベルに応じた処理を行うことができる。復元画像内の全ての画素について誤差拡散処理が施された後、ステップＳ１５〜ステップＳ１８からステップＳ１９に進み、図４のフローチャートが完了する。
【００９０】
以上説明したように、図１の画像処理装置３は、具体的には、画像の符号化データを復号化することによって得られたデータに対して、ブロック毎の分散を求めることによって、ブロックノイズおよびリンギングノイズに代表されるノイズの発生の有無、該ノイズの種類、およびノイズのレベルを求める。さらに、画像処理装置３では、求められた各ノイズに応じたノイズ除去処理を行い、さらに、該ノイズに応じた適切な中間調処理を行う。これによって、画像処理装置３は、画像の復号化データからノイズを効果的に除去することができる。
【００９１】
次に、上述した画像処理方法を実行する画像処理装置３のより詳細な構成について、図１に基づいて説明する。図１の画像処理装置３は、復元処理部２０と階調再現処理部８と空間フィルタ処理部９との他に、入力階調補正部４１、色空間変換部４２、色補正部４３、黒生成下色除去部４４、出力階調補正部４５、および画像記憶部４６を含む。画像形成装置１は、画像処理装置３と画像出力装置４との他に、操作パネル４８を含む。画像処理装置３と画像出力装置４とを含む画像形成装置１は、たとえば、所謂複合機またはプリンタによって実現される。操作パネル４８は、画像形成装置１の利用者によって操作され、操作結果が画像処理装置３および画像出力装置４に与えられ、適宜処理される。
【００９２】
画像処理装置３には、各種の画像入力装置で生成された画像のデータが、インターフェイスを介して入力されている。与えられる画像のデータとしては、たとえば、デジタルカメラで撮像された画像の符号化データ、およびコンピュータ上でソフトウェアを用いて動画からキャプチャされた静止画像のデータが挙げられる。画像入力装置と画像処理装置３との間のインターフェイスの規格としては、ＵＳＢ（Universal Serial Bus）およびＩＥＥＥ１３９４（Institute of Electrical and Electronic Engineers：米国電気電子学会）が挙げられる。画像処理装置３に与えられる画像の符号化データは、たとえばＪＰＥＧに代表される規格のデータとしてフォーマットされている。
【００９３】
画像処理装置３に与えられた画像のデータは、画像処理装置３内において、復元処理部２０、入力階調補正部４１、空間フィルタ処理部９、色空間変換部４２、色補正部４３、黒生成下色除去部４４、出力階調補正部４５、および階調再現処理部８の順で順次送られる。画像処理装置３に与えられた画像のデータは、最初はＹＣｂＣｒ信号形態のデジタルカラー信号であり、最終的に、ＣＭＹＫ信号形態のデジタルカラー信号である画像のデータとして、画像処理装置３から画像出力装置４へ出力される。
【００９４】
復元処理部２０は、ＹＣｂＣｒ信号形態の画像の符号化データを復号化し、逆量子化処理および逆直交変換処理をさらに行うことによって、ＹＣｂＣｒ信号形態の画像の符号化データをＹＣｂＣｒ信号形態の該画像の復号化データに変換する。さらに復元処理部２０は、得られた復号化データが示す復元画像にブロックノイズおよびリンギングノイズが生じているか否かを判定し、ノイズの発生状況の判定結果をノイズ判定信号として出力する。
【００９５】
入力階調補正部４１は、復号処理部６で復号化されたＹＣｂＣｒ信号方式の画像の復号化データの画素のデータに対して、トーンスケールの補正処理およびカラーバランスの補正処理を、必要に応じて施す。空間フィルタ処理部９は、入力階調補正部４１より入力されるＹＣｂＣｒ信号の画像のデータに対して、ノイズ判定信号を基にデジタルフィルタによる空間フィルタ処理を行い、空間周波数特性を補正することによってブロックノイズやリンギングノイズを防ぐように処理するものである。
【００９６】
色空間変換部４２は、ＹＣｂＣｒ信号形態の画像の復号化データの色空間をＬ^＊ａ^＊ｂ^＊表色系の色空間に変換する。色空間変換後のＬ^＊ａ^＊ｂ^＊信号形態の画像の復号化データが、後段の色補正部４３に入力される。色補正部４３は、与えられたＬ^＊ａ^＊ｂ^＊信号形態の画像の復号化データに対して、画像出力装置４の特性に合わせた色補正を施し、色空間変換後のＣＭＹ信号形態の画像の復号化データに変換する。色補正後のＣＭＹ信号形態の画像の復号化データが、後段の黒生成下色除去部４４に入力される。
【００９７】
黒生成下色除去部４４は、色補正部４３から与えられるＣＭＹ信号形態の画像のデータに基づき、黒生成処理と下色除去処理とを行い、ＣＭＹ信号形態の画像のデータをＣＭＹＫ信号形態の画像のデータに変換する。黒生成処理は、ＣＭＹ信号形態の画像のデータに含まれるＣＭＹの３色の色信号から黒信号を生成する処理である。下色除去処理は、ＣＭＹ信号形態の画像のデータに含まれるＣＭＹの３色の色信号から黒の色信号分を差引いて新たなＣＭＹの色信号を生成する処理である。ＣＭＹＫ信号形態の画像のデータは、画素毎に、黒生成処理で生成された黒の色信号と下色除去処理で生成されたＣＭＹの色信号とを含む。
【００９８】
黒生成処理の一例として、スケルトンブラックを用いて黒生成を行なう手法がある。スケルトンブラックを用いる手法では、被処理画素に対する黒生成処理および下色除去処理が同時に行われる。前記手法では、具体的には、被処理画素の色信号が、以下の連立式で示すように変換される。以下の連立式では、スケルトンカーブの入出力特性を「ｙ＝ｆ(x)」、入力されるＣＭＹ信号形態の画素の各色信号を「Ｃ」，「Ｍ」，「Ｙ」、出力されるＣＭＹＫ信号形態の画素の各色信号を「Ｃ＃」，「Ｍ＃」，「Ｙ＃」，「Ｋ＃」、ＵＣＲ（Under Color Removal）率を「α」（０＜α＜１）と表す。ＵＣＲ率αは、０より大きく１未満の数値である。

【００９９】
出力階調補正部４５は、与えられたＣＭＹＫ信号形態の画像の復号化データに対して、画像出力装置４の階調再現特性に応じた出力階調補正処理を施す。出力階調補正処理は、たとえば、画素毎の色の濃度を示す信号から成る画像のデータを、網点面積率を示す信号からなる画像のデータに変換する処理である。
【０１００】
階調再現処理部８は、出力階調補正後のＣＭＹＫ信号形態の画像の復号化データに対して、いわゆる中間調生成処理である中間調処理を施す。さらに、階調再現処理部８は、ＣＭＹＫ信号形態の画像の復号化データに対して、空間フィルタ処理部９と同様に、ノイズ判定信号に基づく上述の誤差拡散処理を行う。中間調処理は、前記復号化データが表す復元画像を画素に分離して、該画像内の階調がそれぞれ再現されるように、画素の濃度を定める処理である。
【０１０１】
復元処理部２０〜階調再現処理部８までの各処理部２０，４１，９，４２〜４５，８における処理が順次施された画像の復号化データは、好ましくは、たとえば、画像記憶部４６に一旦記憶され、所定のタイミングで読出されて、画像出力装置４に入力される。画像出力装置４は、与えられた画像の復号化データに基づき、該画像を顕像化する。画像出力装置４は、たとえば、電子写真方式およびインクジェット方式に代表される処理方式を用いる印刷装置によって実現される。印刷装置は、顕像化された該画像を、紙等で実現される記録媒体上に出力する。また画像出力装置４は、記録媒体への印刷装置に限らず、他の形態の画像出力装置４で実現されてもよい。たとえば、画像出力装置４は、表示装置によって実現されてもよい。陰極線管(Cathode Ray Tube:CRT)装置および液晶ディスプレイに代表される表示装置は、顕像化された該画像を、自己が備える画面上に表示させる。この場合、画像処理装置３において、黒生成下色除去部４４は不要であるので省略される。
【０１０２】
図９は、本発明の他の実施の形態である画像形成装置６１の構成を示すブロック図である。図１では、復号処理部６とノイズ判定部７と階調再現処理部８とを画像処理装置３が含む構成を説明している。図９の画像形成装置６１では、前述の復号処理部６とノイズ判定部７と階調再現処理部８との機能とをコンピュータ６３に実現させるための画像処理プログラム６４を予め用意し、該画像処理プログラム６４を汎用的なコンピュータ６３にインストールして実行させることによって、画像処理装置３が実現されている。これによって、コンピュータ６３は、画像の符号化データの復号化後に、復元画像内のノイズの発生状況を判定し、該ノイズの発生状況に応じた中間調処理を画像の復号化データに施す。これによって、画像処理プログラム６４は、汎用的なコンピュータ６３を用いて、圧縮符号化されたデータの復号化によって生じたノイズを、復号化されたデータから効果的に除去することができる。
【０１０３】
図９の構成の画像形成装置６１は、具体的には、コンピュータ６３が、アプリケーションソフトウエア７１、プリンタドライバ７２、通信ポートドライバ７３、通信ポート７４、中央演算処理装置(Central Processing Unit:CPU)７５、および記憶部７６を含む。図９の例では、画像処理プログラム６４は、アプリケーションソフトウエア７１内のプログラムとプリンタドライバ７２内のプログラムとに分割されている。通信ポート７４は、たとえば、ＲＳ２３２Ｃ規格に基づく通信ポート７４、またはＬＡＮ用の通信ポート７４で実現される。アプリケーションソフトウエア７１とプリンタドライバ７２と通信ポート７４ドライバ７３とは、たとえば記憶部７６に記憶される。中央演算処理装置７５が、アプリケーションソフトウエア７１とプリンタドライバ７２と通信ポート７４ドライバ７３とに含まれるプログラムを実行する。
【０１０４】
デジタルカメラで撮像された画像のデータおよび動画からキャプチャされた静止画像のデータに代表される画像の符号化データは、コンピュータ６３に与えられる。図９の例では、与えられた画像の符号化データに対して、ノイズ判定処理およびノイズ除去処理がアプリケーションソフトウェアによって実行され、処理が施された復元画像がプリンタから出力される。
【０１０５】
アプリケーションソフトウエア７１のプログラムは、図１の復号処理部６と図１のノイズ判定部７との機能をコンピュータ６３に実行させる。好ましくは、アプリケーションソフトウエア７１のプログラムは、図１の空間フィルタ処理部９の機能をコンピュータ６３にさらに実行させる。これらの結果、コンピュータ６３に与えられた画像の符号化データに対して復号化処理が施され、得られた画像の復号化データに対して、復元画像のノイズ判定処理とノイズ除去用の空間フィルタ処理とが、コンピュータ６３によって施される。
【０１０６】
プリンタドライバ７２のプログラムは、概略的には、空間フィルタ処理後の画像の復号化データに対するノイズ除去処理を含む中間調処理を、コンピュータ６３に実行させる。具体的には、プリンタドライバ７２のプログラムを実行するコンピュータ６３は、空間フィルタ処理後の画像のデータに対して、図１の色補正部４３の処理と図１の黒生成下色除去部４４の処理と図１の出力階調補正部４５の処理と図１の階調再現処理部８の処理とを施す。プリンタドライバ７２のプログラムを実行するコンピュータ６３は、さらに、プリンタ言語翻訳部７７を実現している。プリンタ言語翻訳部７７は、中間調処理後の画像の復号化データをプリンタ言語に変換する。プリンタ言語変換後の画像の復号化データは、通信ポート７４ドライバ７３および通信ポート７４を介して、画像出力装置４であるプリンタに与えられる。
【０１０７】
また上述の画像処理プログラム６４は、コンピュータ読取り可能な記録媒体に記録されていることが好ましい。このような記録媒体を用いれば、画像の符号化データの復号化によって生じたノイズを画像の復号化データから効果的に除去することが可能な画像処理プログラム６４を、汎用的なコンピュータ６３に容易に供給することが可能になる。これによって、前記画像処理プログラム６４を記録した記録媒体が持ち運び自在に提供可能になるので、前記画像処理プログラム６４によって実現される機能を容易に頒布することができる。
【０１０８】
本実施の形態では、具体的には、コンピュータ６３内の中央演算処理装置７５が画像処理プログラム６４を実行している。画像処理プログラム６４を格納している記録媒体であるプログラムメディアとして、たとえば、コンピュータ６３の記憶部７６に用いられる不揮発性メモリそのものが用いられてもよい。または、コンピュータ６３の記憶部７６が外部記憶装置として媒体読取り装置を設けているならば、該媒体読取り装置に着脱自在に挿入可能であって、かつ該媒体読取り装置に挿入することで内部のプログラムをコンピュータ６３に読取らせることが可能な記録媒体を、前記プログラムメディアとして用いてもよい。
【０１０９】
中央演算処理装置７５は、記録媒体に格納されている画像処理プログラム６４は、たとえば、コンピュータ６３内の中央演算処理装置７５が該記録媒体に直接アクセスして実行される構成を有する。または、記録媒体に格納されている画像処理プログラム６４は、記録媒体から一旦読出され、記憶部７６内のプログラム記憶エリアにダウンロードされて、ダウンロードされた該画像処理プログラム６４が中央演算処理装置７５に実行される構成であってもよい。プログラムメディアが上述の２種類の形態のいずれであっても、画像処理プログラム６４の構成は、上述の２種類のどちらでもよい。なお、プログラムのダウンロードのためのプログラムは、コンピュータ６３に予め格納されているものとする。
【０１１０】
画像処理プログラム６４を格納している記録媒体であるプログラムメディアは、好ましくは、コンピュータ６３本体と分離可能に構成されている。プログラムメディアとして用いられる記録媒体としては、たとえば、テープ系記録媒体、ディスク系記録媒体、カード系記録媒体、およびメモリ系記録媒体が挙げられる。テープ系記録媒体としては、たとえば、磁気テープおよびカセットテープが挙げられる。ディスク系記録媒体としては、たとえば、フロッピー（Ｒ）ディスクおよびハードディスク装置に代表される磁気ディスク媒体、ＣＤ−ＲＯＭ（Compact Disc−Read Only Memory）およびＤＶＤ（Digital Video Disc）に代表される光ディスク媒体、ならびに、ＭＯ（Magneto-Optic Disc）およびＭＤ(Mini Disc)に代表される光磁気ディスク媒体が挙げられる。カード系記録媒体としては、所謂メモリカードを含むＩＣ(Integrated Circuit)カードおよび光カードが挙げられる。メモリ系記録媒体は、たとえば、不揮発性の半導体メモリが挙げられる。不揮発性の半導体メモリは、たとえば、マスクＲＯＭ、ＥＰＲＯＭ（Erasable Programmable Read Only Memory）、ＥＥＰＲＯＭ（Electrically Erasable Programmable Read Only Memory）、およびフラッシュＲＯＭを含む。このように、プログラムメディアに用いられる記録媒体は、固定的にプログラムを担持する媒体であればよい。
【０１１１】
また上述した画像処理プログラム６４は、記録媒体を介してコンピュータ６３に与えられる状況に限らず、伝送媒体を介した通信によってコンピュータ６３に与えられてもよい。伝送媒体としては、電話回線網の伝送路、ＬＡＮ（Local Area Network）およびＷＡＮ（Wide Area Network）等を用いたコンピュータネットワークの伝送路、およびコンピュータを接続する伝送ケーブル等が挙げられる。これによって、画像処理装置３の実現がより容易になる。具体的には、図９の画像処理装置３がインターネットを含む通信ネットワークに接続可能な構成を有する場合、通信ネットワークから画像処理プログラム６４がダウンロードされてもよい。通信ネットワークから画像処理プログラム６４がダウンロードされる場合、ダウンロード用のプログラムは、コンピュータ６３本体に予め格納しておいてもよく、または別の記録媒体からコンピュータ６３に事前にインストールされればよい。
【０１１２】
以上説明したように、図１および図９の画像処理装置３は、具体的には、圧縮符号化された画像のデータを復号化することによって得られたデータに対して、ブロック毎の分散を求めることによって、ブロックノイズおよびリンギングノイズに代表されるノイズの発生の有無、該ノイズの種類、およびノイズのレベルが求められる。さらに、画像処理装置３では、求められた各ノイズに応じたノイズ除去処理を行い、さらに、該ノイズに応じた適切な中間調処理を行う。
【０１１３】
第１ならびに第２の実施の形態の画像形成装置１，６１および画像処理装置３は、本発明の画像形成装置および画像処理装置３の例示であり、主要な構成および動作が等しければ、他の様々な形で実現することができる。特に画像処理装置３の各構成部品の詳細な構成および動作は、同じ効果が得られるならば、上述の構成および動作に限らず、他の構成および動作によって実現されてもよい。画像処理装置３の処理部は、個別の回路によってそれぞれ実現される構成に限らず、１または複数の処理部の機能を実現するための制御ソフトウェアをコンピュータに実行させることによって、機能的に実現されてもよい。
【０１１４】
【発明の効果】
以上のように本発明によれば、画像処理装置は、圧縮符号化された画像のデータを復号化し、復号化されたデータが表す画像内のノイズの発生状況をブロック単位で判定して、判定結果に基づいて、階調を再現させるだけでなくノイズを低減させるような中間調処理を該復号化されたデータに施す。これによって、画像処理装置は、中間調処理を行いつつ、画像の復号化されたデータ内のノイズを効果的に除去することができる。
【０１１５】
また本発明によれば、画像処理装置において、復号化されたデータが表す画像が１以上のブロックに分割され、中間調処理がブロック毎に施される。これによって、画像処理装置は、中間調処理後の画像内のブロックの境界でのテクスチャの発生を防止することができる。さらにまた本発明によれば、画像処理装置において、中間調処理として、画像の復号化されたデータに対する誤差拡散処理が行われる。誤差拡散処理で用いられる量子化閾値と誤差判定閾値と拡散係数との３つの定数のうちの少なくとも１つが、画像内のノイズの発生状況の判定結果に応じて切換えられる。これによって、画像処理装置は、中間調処理後の画像内のブロックの境界でのテクスチャの発生を、容易に防止することができる。
【０１１６】
また本発明によれば、画像処理装置において、階調再現処理部の処理に先だって、ノイズを低減させるための空間フィルタ処理が、画像の復号化されたデータに対して施される。これによって、画像処理装置は、復号化されたデータが表す画像からノイズをより効果的に除去することができる。
【０１１７】
さらにまた以上のように本発明によれば、画像形成装置は、画像のブロック毎のノイズの発生状況に基づいて該画像の復号化されたデータに中間調処理を施し、中間調処理が施された画像のデータに基づいて画像を顕像化する。これによって、画像形成装置は、品質のよい出力画像を提供することが可能になる。
【０１１８】
また以上のように本発明によれば、画像処理方法は、圧縮符号化された画像のデータの復号化処理と、復号化されたデータが表す画像内のノイズの発生状況をブロック単位で判定するノイズ判定処理とを行い、ノイズ判定処理の判定結果に基づいて、階調を再現させるだけでなくノイズを低減させるような中間調処理を該画像の復号化されたデータに施す。これによって、画像処理方法が適用される場合、圧縮符号化されたデータの復号化によって生じたノイズを、復号化されたデータから効果的に除去することができる。
【０１１９】
さらにまた以上のように本発明によれば、画像処理プログラムは、画像の圧縮符号化されたデータを復号化する機能と、復号化されたデータが表す画像内のノイズの発生状況をブロック単位で判定する機能と、階調を再現させつつノイズを低減させるための中間調処理を画像の復号化されたデータに対して施す機能とを、コンピュータに実現させる。これによって、画像処理プログラムは、汎用的なコンピュータを用いて、圧縮符号化されたデータの復号化によって生じたノイズを、復号化されたデータから効果的に除去することができる。また以上のように本発明によれば、記録媒体は、上述の画像処理プログラムを記録しており、該画像処理プログラムをコンピュータに読取らせることが可能である。これによって、前記画像処理プログラムによって実現される機能の頒布が容易になる。
【図面の簡単な説明】
【図１】本発明の実施の一形態である画像形成装置１の全体的な構成を示すブロック図である。
【図２】図１の画像処理装置３の復元処理部２０の機能的構成を示すブロック図である。
【図３】図１の画像処理装置３の階調再現処理部８の機能的構成を示すブロック図である。
【図４】図１の画像処理装置３において行われる画像処理を説明するためのフローチャートである。
【図５】図１の画像処理装置のノイズ判定部７において行われるノイズ判定処理で用いられる３×３ブロック群の構成を示す模式図である。
【図６】図１の画像処理装置の空間フィルタ処理部９で用いられるフィルタの構成を示す模式図である。
【図７】図１の画像処理装置の階調再現処理部８において行われる誤差拡散処理で処理対象となる領域を示す模式図である。
【図８】図１の画像処理装置の階調再現処理部８において行われる誤差拡散処理で用いられる誤差拡散範囲を示す模式図である。
【図９】本発明の実施の他の形態である画像形成装置６１の全体的な構成を示すブロック図である。
【符号の説明】
１，６１ 画像形成装置
３ 画像処理装置
４ 画像出力装置
６ 復号処理部
７ ノイズ判定部
８ 階調再現処理部
９ 空間フィルタ処理部
１１ 加算器
１２ 量子化処理部
１３ 誤差判定部
１４ 拡散係数設定部
６３ コンピュータ
６４ 画像処理プログラム
ＴＨａ 量子化閾値
ＴＨｂ 誤差判定閾値
ＴＨｃ 拡散係数
Ｂ５ 注目ブロック
Ｂ１〜Ｂ４，Ｂ６〜Ｂ９ 周辺ブロック[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing apparatus, an image forming apparatus, an image processing method, and an image processing program configured to perform image processing including decoding of image compression-encoded data and halftone processing on the decoded data. And a recording medium.
[0002]
[Prior art]
In recent years, due to the progress of digitalization of office automation (OA) equipment, the spread of computers, and the development of communication networks, the acquisition of image data from an image input device to a computer and the network The transmission and reception of image data between computers via the computer can be performed relatively easily. The image input device is represented by a scanner, for example. Image data given to an image processing device typified by a computer and OA equipment is stored in a recording medium typified by a hard disk device, a DVD-RAM (Digital Versatile Disc-Random Access Memory), or the like. The image data in the image processing apparatus is visualized and output by an image output apparatus typified by a color printer, if necessary. Examples of the color printer system include an electrophotographic system, an inkjet system, and a thermal transfer system.
[0003]
Image data generally has a larger amount of information and a larger file size than document data. For this reason, various methods for compressing image data while suppressing deterioration in image quality have been proposed. Currently, as one of the most widely used image data compression methods, there is a block coding method using a discrete cosine transform (DCT).
[0004]
In the block coding method, first, an image is divided into a plurality of blocks each having a size of 8 pixels vertically and horizontally (8 × 8 pixels), for example. Next, using a discrete cosine transform method that is a two-dimensional orthogonal transform, the intensity of the signal indicating the density distribution in each block is transformed into a DCT coefficient that is a spatial frequency component. Subsequently, the DCT coefficient is quantized for each block, and then an encoded code is assigned to the quantized DCT coefficient using, for example, variable length coding represented by Huffman coding.
[0005]
When compressing the image data, the DCT coefficient corresponding to the high frequency component of the original image is cut off from the image data and the data is compressed, and the DCT coefficient of the high frequency component of the image represented by the decoded data is compressed. It is difficult to visually determine the presence or absence of truncation. Utilizing this phenomenon, the DCT coefficient quantization process quantizes the DCT coefficient of the high frequency component by dividing it by a larger value than the DCT coefficient of the low frequency component, thereby reducing the amount of information. is doing.
[0006]
The image data encoded using the above-described block encoding method is subjected to a decoding process that is the reverse of the block encoding method in the decoding side device, whereby the original image data is obtained. Is restored. In other words, the decoding-side device performs decoding processing, inverse quantization processing, and inverse orthogonal transform processing on the encoded image data, thereby obtaining image-decoded data. However, when decoding, an error due to quantization processing increases, that is, if the difference in DCT coefficients between the image before encoding and the original image restored by quantization and inverse quantization increases, Even if the data is decoded, it is difficult to correctly restore the original image. That is, compared to the original image before encoding, the image represented by the decoded data is distorted, and the occurrence of noise that is visually noticeable is seen. Specifically, for example, noise represented by block noise and ringing noise is generated in an image represented by decoded data. Block noise is noise that makes an image appear to be divided into blocks. Ringing noise is also referred to as mosquito noise, and is foggy or ringing noise that appears around the contour of an image.
[0007]
In a conventional image processing apparatus, in order to reduce noise typified by block noise and ringing noise from within an image represented by decoded data, smoothing processing is performed on the decoded image data. . The types and levels of distortions that occur in decoded image data are generally diverse. For this reason, when the smoothing process is performed in accordance with the largest distortion of the decoded image data, the image data is blurred, and in some cases, the image quality of the image after the smoothing process is greatly reduced. If smoothing processing is performed in accordance with intermediate distortion of image data, distortion remains in the image data, and the image quality is not sufficiently improved.
[0008]
In order to reduce noise from the image represented by the decoded data, the following first conventional technique has been proposed (Patent Document 1). In the first conventional image signal decoding method, distortion is removed using a smoothing process in accordance with the type and level of distortion generated in an image represented by decoded data. Specifically, in the first conventional technique, when decoding compressed image data, a plurality of variables for determining distortion are created based on orthogonal transform coefficients for each block. Based on the plurality of variables, the presence / absence of distortion, the type of distortion, and the level of distortion, that is, the degree of distortion, are determined, and further, noise is removed by removing distortion for each block based on the determination result of distortion. Has been reduced. That is, in the first prior art, the type of distortion and the degree of distortion of each block in the image are analyzed, and by performing smoothing processing based on the analysis result, the data of the decoded image is analyzed. Distortion is reduced.
[0009]
[Patent Document 1]
JP 2001-346208 A
[0010]
[Problems to be solved by the invention]
When the first prior art described above is used, noise in the image represented by the decoded data is reduced using a smoothing process. However, the first prior art makes no consideration regarding halftone processing performed on image data when the image data is transferred from the image processing apparatus to the image output apparatus. For this reason, when data of an image processed using the first conventional technique is given to an image output device to visualize and output the image, blur and noise are generated in the output image. is there. That is, even in the case of image data decoded by the first conventional technique, when the image is output by a technique such as printout, it is difficult to completely remove the distortion in the image.
[0011]
The present invention relates to an image processing apparatus and an image processing method capable of removing noise from an image indicated by decoded data while performing decoding processing and halftone processing on compression-encoded data of the image An image forming apparatus, an image processing program, and a recording medium are provided.
[0012]
[Means for Solving the Problems]
The present invention is provided with a decoding processing unit that is provided with compression-encoded data representing an image and decodes the compression-encoded data.
The image represented by the decoded data contains multiple pixels The target block and the surrounding blocks around the target block Divided into By calculating the variance of the pixel values of all pixels for each block and performing threshold processing based on the calculated variance Determines whether or not noise is generated in the block for each block, and determines the type of noise and the degree of occurrence of noise for blocks with noise And selecting a noise removal filter to be used for the spatial filter processing based on the determination result A noise determination unit;
A gradation reproduction processing unit that performs halftone processing for reducing the noise while reproducing the gradation based on the determination result of the noise determination unit, on the decoded data of the image, An image processing apparatus.
[0013]
According to the present invention, the image processing apparatus includes a decoding processing unit that decodes compression-encoded image data and a gradation reproduction processing unit that performs halftone processing on the decoded image data. In addition, a noise determination unit that determines the occurrence of noise in the image represented by the decoded data in units of blocks. The noise determination unit divides the block of interest into peripheral blocks around the block of interest, calculates the variance of pixel values of all pixels for each block, and performs threshold processing based on the calculated variance. Whether or not noise is generated in the block is determined for each block by threshold processing, and for the block in which noise is generated, the type of the noise and the degree of occurrence of the noise are determined, and based on the determination result The noise removal filter used for the spatial filter processing is selected. The halftone process is adaptively executed based on the determination result of the noise determination unit so as to not only reproduce the gradation but also reduce the noise. Thus, the image processing apparatus effectively performs the halftone process on the decoded data of the image, and effectively eliminates the noise generated by the decoding of the compression encoded data from the decoded data. Can be removed.
[0014]
In the image processing apparatus of the present invention, the gradation reproduction processing unit divides the image represented by the decoded data into one or more blocks including a plurality of pixels, and performs the halftone processing for each block. It is characterized by.
[0015]
According to the present invention, in the image processing apparatus, the image represented by the decoded data is divided into one or more blocks, and a halftone process different for each block is performed in the gradation reproduction processing unit. As a result, it is possible to adaptively change parameters used for halftone processing according to the type of noise generated in each block and the degree of noise. As a result, the image processing apparatus can perform halftone processing without generating a texture at a block boundary.
[0016]
In the image processing apparatus of the present invention, the gradation reproduction processing unit performs halftone processing by performing error diffusion processing on the decoded data of the image.
(A) based on the decoded data of the image, for each pixel constituting the image, an adder that obtains a diffusion pixel value that is the sum of a pixel value and a diffusion error;
(B) By quantizing the diffusion pixel value of each pixel based on a predetermined quantization threshold, a quantization value to be the pixel value of the halftone processed image data is obtained for each pixel, and diffusion is performed A quantization processing unit that further obtains a quantization error, which is a difference between the pixel value and the quantization value, for each pixel;
(C) an error determination unit that corrects a quantization error based on a predetermined error determination threshold;
(D) a diffusion coefficient setting unit that calculates a diffusion error for allocating a corrected quantization error to an unprocessed pixel value based on a predetermined diffusion coefficient;
At least one of the quantization threshold, the error determination threshold, and the diffusion coefficient is switched based on a determination result of the noise determination unit.
[0017]
According to the present invention, in the image processing apparatus, error diffusion processing is performed on the decoded data of the image as halftone processing. At least one of the three constants of the quantization threshold, the error determination threshold, and the diffusion coefficient used in the error diffusion process is switched according to the determination result of the noise determination unit. As a result, it is possible to adaptively change the error diffusion process according to the noise generation state in the image. That is, it is possible to perform error diffusion processing that not only reproduces gradation but also reduces noise by a simple method of switching at least one of the above three constants. Thus, the image processing apparatus can easily perform halftone processing that can remove noise from the decoded data without generating a texture at a block boundary.
[0018]
The image processing apparatus of the present invention further includes a spatial filter processing unit that performs spatial filter processing for reducing noise on the decoded data of the image based on the determination result of the noise determination unit,
The gradation reproduction processing unit performs halftone processing on image data that has been subjected to spatial filter processing.
[0019]
According to the present invention, in the image processing apparatus, prior to the processing of the gradation reproduction processing unit, spatial filter processing for reducing noise is performed on the decoded data of the image. As a result, not only processing for reducing noise is performed in the spatial filter processing unit, but also halftone processing is adaptively performed in the tone reproduction processing unit so as to reduce noise. Thus, noise is reduced in both the spatial filter processing unit and the gradation reproduction processing unit, so that the image processing apparatus can more effectively remove noise from the image represented by the decoded data.
[0020]
The present invention also provides the above-described image processing apparatus,
An image forming apparatus comprising: an image output device which receives image data subjected to halftone processing and visualizes the image based on the image data.
[0021]
According to the present invention, the image forming apparatus includes an image processing apparatus and an image output apparatus. In the image processing apparatus, halftone processing is performed on the decoded data of the image based on the noise generation state for each block of the image represented by the decoded data. As a result, it is possible to reduce blurring and noise that occur in the output image when the image is visualized and output by the image output apparatus. As a result, the image forming apparatus can provide a high-quality output image.
[0022]
The present invention also includes a process for decoding compression-coded data representing an image;
The image represented by the decoded data contains multiple pixels The target block and the surrounding blocks around the target block Divided into By calculating the variance of the pixel values of all pixels for each block and performing threshold processing based on the calculated variance Determines whether or not noise is generated in the block for each block, and determines the type of noise and the degree of occurrence of noise for blocks with noise And selecting a noise removal filter to be used for the spatial filter processing based on the determination result Processing,
An image processing method comprising: performing halftone processing for reducing noise while reproducing gradation based on a determination result of noise on the decoded data of the image .
[0023]
According to the present invention, in the image processing method, in addition to the decoding process of the compression-encoded image data and the halftone process for performing the halftone process on the decoded image data, the decoded image data is decoded. It includes a noise determination process for determining the occurrence of noise in an image represented by data in units of blocks. The noise determination process is divided into a target block and peripheral blocks around the target block, calculates a variance of pixel values of all pixels for each block, and performs threshold processing based on the calculated variance. Whether or not noise is generated in the block is determined for each block by threshold processing, and for the block in which noise is generated, the type of the noise and the degree of occurrence of the noise are determined, and based on the determination result The noise removal filter used for the spatial filter processing is selected. The halftone process is adaptively executed based on the determination result of the noise determination process so that not only the gradation is reproduced but also the noise is reduced. Thereby, when the image processing method is applied, noise generated by decoding the compression-encoded data can be effectively removed from the decoded data.
[0024]
The present invention also includes a function of decoding compression-coded data representing an image,
The image represented by the decoded data contains multiple pixels The target block and the surrounding blocks around the target block Divided into By calculating the variance of the pixel values of all pixels for each block and performing threshold processing based on the calculated variance Determines whether or not noise is generated in the block for each block, and determines the type of noise and the degree of occurrence of noise for blocks with noise And selecting a noise removal filter to be used for the spatial filter processing based on the determination result Function and
An image characterized by causing a computer to realize a function of performing halftone processing for reducing noise while reproducing gradation based on a determination result of noise on the decoded data of the image It is a processing program.
[0025]
According to the present invention, the image processing program is executed by a computer. As a result, after decoding the compression-encoded image data, the computer determines the occurrence of noise in the image represented by the decoded data, and performs halftone processing according to the occurrence of the noise. To the data. In determining the noise occurrence state, the block is divided into the block of interest and peripheral blocks around the block of interest, the variance of pixel values of all pixels is calculated for each block, and threshold processing is performed based on the calculated variance. Whether or not noise is generated in the block is determined for each block by threshold processing, and for the block in which noise is generated, the type of the noise and the degree of occurrence of the noise are determined, and based on the determination result The noise removal filter used for the spatial filter processing is selected. Thus, the image processing program can effectively remove noise generated by decoding the compression-encoded data from the decoded data using a general-purpose computer.
[0026]
The present invention also provides a computer-readable recording medium on which the above-described image processing program is recorded.
[0027]
According to the present invention, the recording medium records an image processing program to be executed by a computer, and the computer can read the image processing program. The image processing program has a function for decoding the compression-coded data of the image, a function for determining the noise occurrence status in the image represented by the decoded data, and a noise while reproducing the gradation. And a function of applying halftone processing to the decoded data of the image to reduce the image quality. If such a recording medium is used, an image processing program capable of effectively removing noise generated by decoding compression-encoded data from the decoded data can be easily obtained on a general-purpose computer. It becomes possible to supply. This facilitates the distribution of functions realized by the image processing program.
[0028]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 is a block diagram showing an overall configuration of an image forming apparatus 1 according to an embodiment of the present invention. The image forming apparatus 1 includes an image processing device 3 and an image output device 4. Compressed and encoded data representing an image composed of a plurality of pixels is given to the image processing apparatus 3 from the outside. Basically, the image processing device 3 performs various processes on the compression-encoded data of the given image, and gives the processed image data to the image output device 4. The image output device 4 visualizes and outputs the image based on the processed image data given from the image processing device 3.
[0029]
The image processing apparatus 3 in FIG. 1 relates to decoding of data that has been compression-encoded for an image. Schematically, the image processing apparatus 3 performs decoding of data that has been compression-encoded for an image, detects noise generated during the decoding, and performs appropriate noise removal processing according to the detection result and A halftone process is performed on the decoded image data. The halftone process is basically a process for reproducing the gradation of each pixel of the image based on the image data. In the following description, the compression-encoded data and the image indicated by the data are abbreviated as “encoded data” and “compressed image”, respectively, and the decoded data and the image indicated by the data are indicated as “decoded data”. And may be abbreviated as “restored image”.
[0030]
The image processing device 3 includes at least a decoding processing unit 6, a noise determination unit 7, and a gradation reproduction processing unit 8. The encoded data of the image is given to the reprocessing unit 6. The reprocessing unit 6 decodes the encoded data of the image and outputs the decoded data of the image. The noise determination unit 7 first divides an image represented by the decoded data into one or more blocks including a plurality of pixels. Next, the noise determination unit 7 determines whether or not noise is generated in the block for each block, and determines the type of noise and the degree of occurrence of the noise regarding the block in which noise is generated. Based on the determination result of the noise determination unit 7, the gradation reproduction processing unit 8 performs halftone processing for reducing the noise while reproducing the gradation on the decoded data of the image. The image data processed by the gradation reproduction processing unit 8 is given to the image output device 4 as output data from the image processing device 3.
[0031]
As described above, the image processing apparatus 3 in FIG. 1 includes the noise determination unit 7 that determines the occurrence of noise in the restored image in units of blocks in addition to the decoding processing unit 6 and the gradation reproduction processing unit 8. The halftone process is adaptively executed based on the determination result of the noise determination unit 7 so as to not only reproduce the gradation but also reduce the noise. As a result, the image processing apparatus 3 performs processing in consideration of noise generated by decoding the encoded data during halftone processing on the decoded data of the image. Can be effectively removed.
[0032]
The image processing apparatus 3 preferably further includes a spatial filter processing unit 9. Prior to the processing of the gradation reproduction processing unit 8, the spatial filter processing unit 9 performs spatial filter processing for reducing noise on the decoded data of the image based on the determination result of the noise determination unit 7. One typical spatial filter process is a smoothing process. As a result, not only a process for reducing noise is performed in the spatial filter processing unit 9, but also a halftone process is adaptively performed in the gradation reproduction processing unit 8 so as to reduce noise. As a result, the image processing device 3 reduces noise in the restored image in both the spatial filter processing and the halftone processing, so that noise can be more effectively removed from the restored image.
[0033]
Furthermore, the gradation reproduction processing unit 8 of the image processing device 3 preferably divides the decoded data of the image into portions corresponding to one or more blocks including a plurality of pixels, and performs the halftone processing for each block. Apply. As a result, it is possible to adaptively change parameters used for halftone processing according to the type of noise generated in each block and the degree of noise. This prevents a texture from being generated at the block boundary due to the halftone process.
[0034]
Further, in the gradation reproduction processing unit 8 in the image processing device 3, for example, specifically, error diffusion processing is performed on decoded image data as halftone processing. For error diffusion processing, the gradation reproduction processing unit 8 includes an adder 11, a quantization processing unit 12, and a diffusion coefficient setting unit 14, and preferably further includes an error determination unit 13. The adder 11 obtains a diffusion pixel value that is the sum of the pixel value and the diffusion error for each pixel constituting the image based on the decoded data of the image. The quantization processing unit 12 quantizes the diffusion pixel value of each pixel based on a preset quantization threshold THa, thereby obtaining a quantization value to be a pixel value of the halftone processed image data. Obtained for each pixel. And the quantization process part 12 calculates | requires the quantization error which is a difference of a diffused pixel value and a quantization value for every pixel further. The error determination unit 13 corrects the quantization error based on a preset error determination threshold value THb. The diffusion coefficient setting unit 14 calculates a diffusion error for allocating the corrected quantization error to an unprocessed pixel value based on a predetermined diffusion coefficient THc.
[0035]
When the gradation reproduction processing unit 8 performs error diffusion processing, preferably at least one of the quantization threshold THa and the diffusion coefficient THc is switched based on the determination result of the noise determination unit 7. When the error determination unit 13 is included, preferably at least one of the three constants of the quantization threshold THa, the error determination threshold THb, and the diffusion coefficient THc is switched based on the determination result of the noise determination unit 7. . As a result, it is possible to adaptively change the error diffusion process according to the noise generation state in the restored image. That is, using a simple method of switching at least one of the three constants described above, it is possible to perform error diffusion processing that not only reproduces gradation but also reduces noise. As a result, the image processing apparatus 3 can easily perform halftone processing that can remove noise from the decoded data without generating a texture at a block boundary.
[0036]
As described above, the image processing apparatus 3 in FIG. 1 performs processing for reducing noise typified by block noise and ringing noise when decoding encoded image data, and also decodes the image. In the halftone process for reproducing the gradation of each pixel of the image for data output, a process for reducing noise is performed. As a result, even when distortion due to noise remains during decoding of the encoded data of the compressed image, the generation of texture due to the distortion during halftone processing is reduced, so the image quality of the output image data Is prevented.
[0037]
An image forming apparatus 1 in FIG. 1 includes the above-described image processing apparatus 3 and an image output apparatus 4 that performs image visualization. Prior to the visualization of the image, the decoded data is subjected to halftone processing for reducing the noise while reproducing the gradation based on the noise generation state for each block of the restored image. As described above, since noise removal is considered in the halftone process, blurring and noise generated in the output image are reduced when the image is visualized and output. As a result, the image forming apparatus 1 can improve the quality of the output image.
[0038]
As a signal form used to represent image data in the image processing apparatus 3, a so-called YCbCr signal form, L ^* a ^* b ^* Examples include signal forms, CMY signal forms, and CMYK signal forms. In the YCbCr signal form, the color is expressed by combining three elements of the luminance signal component Y and the two types of color difference signal components Cr and Cb. In the CMY signal form, a color is expressed by combining three subtractive color elements of cyan, magenta, and yellow. The CMYK signal form is widely used in the field of printing, and a color is expressed by combining four color elements obtained by adding black to three subtractive colors. L ^* a ^* b ^* The signal form is so-called L ^* a ^* b ^* Color space is used. L ^* a ^* b ^* The color space is a color space of the color system standardized by the Commission Internationale de l'Eclairage (CIE) in 1976. ^* And two chromaticities a indicating hue and saturation, respectively ^* , B ^* The color is expressed using the three elements. In the following description, a combination of three colors of cyan, magenta, and yellow is abbreviated as “CMY”, and a combination of four colors of cyan, magenta, yellow, and black (black) is abbreviated as “CMYK”.
[0039]
FIG. 2 is a block diagram illustrating a functional configuration of the restoration processing unit 20 of the image processing apparatus 3 of FIG. The decoding processing unit 6 and the noise determination unit 7 are included in the restoration processing unit 20. FIG. 3 is a block diagram showing a functional configuration of the gradation reproduction processing unit 8 of the image processing apparatus 3 of FIG. A detailed configuration of the image processing apparatus 3 in FIG. 1 will be described below with reference to FIGS. In the following example, the JPEG (Joint Photographic Experts Group) format is adopted as the data format of the image data, and the Huffman coding method is adopted as the encoding method of the image data in the JPEG format. In the following example, encoded data of an image in the form of a YCbCr signal is given from an image input device typified by a digital camera to the image processing device 3.
[0040]
The restoration processing unit 20 preferably further includes a parameter calculation unit 22 and a threshold storage unit 23 in addition to the decoding processing unit 6 and the noise determination unit 7. Specifically, the decoding processing unit 6 includes a variable length code decoding unit 25, an inverse quantization unit 26, and an inverse orthogonal transform unit 27. In addition to the adder 11, the quantization processing unit 12, the error determination unit 13, and the diffusion coefficient setting unit 14, the gradation reproduction processing unit 8 is preferably a quantization threshold storage unit 31, a diffusion coefficient storage unit 32, an error A storage unit 33 and a density determination unit 34 are further included. The error determination unit 13 may be omitted.
[0041]
The encoded data of the image in the YCbCr signal format is given to the variable length code decoding unit 25. The variable length code decoding unit 25 decodes encoded data of a given image. By the decoding process of the variable-length code decoding unit 25, the quantized DCT coefficients are separated from the encoded data of the image. The inverse quantization unit 26 restores the DCT coefficient by inversely quantizing the quantized DCT coefficient. Further, the inverse orthogonal transform unit 27 performs inverse orthogonal transform on the restored DCT coefficient, thereby transforming the restored DCT coefficient into pixel data. The obtained pixel data indicates the color and density of the pixel of the restored image, and in the example of FIG. 2 is data in the form of a YCbCr signal. A set of all obtained pixel data is transmitted to the input gradation correction unit 41 as decoded data of an image in the YCbCr signal format. Also, only the luminance signal component Y in the decoded data of the image in the YCbCr signal format is transmitted to the parameter calculation unit 22.
[0042]
The parameter calculation unit 22 calculates a noise determination parameter using the luminance signal component Y in the decoded image data obtained by the inverse quantization unit 26. The noise determination parameter is a parameter for determining the presence or absence of noise in the restored image. The noise to be determined is, for example, block noise and ringing noise. In the example of FIG. 2, the presence or absence of noise is determined by the parameter calculation unit 22 using the luminance signal component Y because the human eye is not the change of the color difference signal components Cb and Cr but the change of the luminance signal component Y. Due to being sensitive. Of course, at least one of the two types of color difference signal components Cb and Cr may be used for noise determination. In the example of FIG. 2, as described later, the variance of pixel data for each block in the restored image is obtained as the noise determination parameter.
[0043]
The noise determination unit 7 compares the noise determination parameter obtained by the parameter calculation unit 22 with a predetermined determination threshold, and determines the occurrence of noise typified by block noise and ringing noise for each block. The determination threshold is stored in the threshold storage unit 23 in advance. For example, the noise occurrence state is specifically defined by the following first to third situation elements. As the first situation element, the presence or absence of block noise and ringing noise is determined. As the second situation element, the level of noise when block noise occurs, that is, the degree of noise in that case is determined. As the third situation element, the level of smoothing when neither block noise nor ringing noise occurs is determined. The determination result of the noise determination unit 7 is output as a noise determination signal. The noise determination signal is given to the quantization processing unit 12, the error determination unit 13, the diffusion coefficient setting unit 14, and the density determination unit 34 provided in the gradation reproduction processing unit 8.
[0044]
In the example of FIGS. 1 to 3, the gradation reproduction processing unit 8 is provided with the decoded data of the image in the CMYK signal format. In the decoded data of the image in the CMYK signal format, the pixel data is represented by a combination of four types of pixel values each indicating the density of each color of CMYK. Each processing unit in the gradation reproduction processing unit 8 individually processes, for example, four types of colors of each pixel. In the following description of FIG. 3, only processing for any one kind of color of any one pixel to be processed will be described.
[0045]
The density determination unit 34 of the gradation reproduction processing unit 8 determines the density level of the color of the pixel to be processed based on the pixel data of the pixel to be processed in the decoded data of the image given to the gradation reproduction processing unit 8. The density determination signal is output to the quantization processing unit 12 so that an appropriate quantization threshold THa is set when performing the determination and the quantization. The adder 11 adds the diffusion error stored in the error storage unit 33 to the pixel value of the pixel to be processed whose density level has been determined by the density determination unit 34. As a result, the diffusion pixel value of the pixel to be processed that is the sum of the pixel value and the diffusion pixel is output.
[0046]
The quantization processing unit 12 compares the given diffusion pixel value of the pixel to be processed with a predetermined quantization threshold THa stored in the quantization threshold storage unit 31 to quantize the diffusion pixel value. Do. As a result, a quantization value to be the pixel value of the pixel to be processed of the halftone processed image data is obtained. The quantized value of the pixel to be processed is output from the image processing device 3 as the pixel value of the pixel to be processed of the halftone processed image data. Further, the quantization processing unit 12 obtains a quantization error of the pixel to be processed that is a difference between the diffusion pixel value of the pixel to be processed and the quantization value of the pixel to be processed.
[0047]
When quantizing the diffusion pixel value, it is preferable to quantize based on the noise determination signal provided from the noise determination unit 7 in the restoration processing unit 20 in FIG. 2 and the density determination signal provided from the density determination unit 34 in FIG. The quantization threshold THa actually used for the pixel to be processed in the quantization processing unit 12 is appropriately switched. For this purpose, the quantization threshold value storage unit 31 may store a plurality of types of quantization threshold values THa, and may sequentially calculate the quantization threshold value THa based on the noise determination signal and the density determination signal.
[0048]
Based on the quantization error of the pixel to be processed obtained by the quantization processing unit 12 and the predetermined diffusion coefficient THc stored in the diffusion coefficient storage unit 32, the diffusion coefficient setting unit 14 A diffusion error indicating the density to be diffused to the pixel value of the pixel to be processed is obtained. The diffusion error obtained by the diffusion coefficient setting unit 14 is stored in the error storage unit 33. The new diffusion error stored in the error storage unit 33 is added to the pixel value of the next new pixel by the adder 11.
[0049]
When calculating the diffusion error, preferably, the diffusion coefficient THc actually used for the pixel to be processed is appropriately switched by the diffusion coefficient setting unit 14 based on the noise determination signal given from the noise determination unit 7. For this purpose, the diffusion coefficient storage unit 32 may store a plurality of types of diffusion coefficients THc, or may sequentially calculate the diffusion coefficient THc based on the noise determination signal.
[0050]
The error determination unit 13 is a processing unit for performing the above-described quantization error correction process in a situation where the density determination unit 34 does not perform the process of saturating the pixel value. When the quantization error is corrected, the diffusion coefficient setting unit 14 obtains the diffusion error based on the corrected quantization error of the pixel to be processed and the diffusion coefficient THc. The error determination unit 13 compares the error obtained by the quantization processing unit 12 with a preset error determination threshold value THb, and determines whether or not to perform correction processing of either the error truncation or rounding. judge. When it is determined that correction processing is to be performed, the quantization error from the quantization processing unit 12 is corrected by the selected correction method, and the determination error that is the corrected quantization error is given to the diffusion coefficient setting unit 14.
[0051]
The determination of the presence or absence of correction in the error determination unit 13 is performed based on the noise determination signal provided from the noise determination unit 7 and the density determination signal provided from the density determination unit 34. For example, if the absolute value of the quantization error before the correction is 20 or less, the quantization error after the correction is corrected to 0, and if the absolute value of the quantization error before the correction is larger than 20, the quantization after the correction is performed. As the error, the value before correction is used as it is. As a result, the situation in which the quantization error is corrected in the error determination unit 13 is the same as the situation in which the pixel determination is performed in the density determination unit 34. For this reason, when the pixel value saturation processing is performed in the density determination unit 34, the error determination unit 13 may be omitted. In the following description, it is assumed that the pixel value saturation process is performed in the density determination unit 34, and the error determination unit 13 will be described as necessary.
[0052]
In the gradation reproduction processing unit 8 of FIG. 3, the detailed processing procedure when the pixel value saturation processing of the density determination unit 34 is not performed is as follows. First, the density determination unit 34 performs density determination based on the noise determination signal for the pixel value of any one pixel in the image data input to the gradation reproduction processing unit 8, and the density determination signal Is output to the quantization processing unit 12. Next, the diffusion error for the pixel stored in the error storage unit 33 is added to the pixel value of the pixel by the adder 11, and the addition result is output to the quantization processing unit 12. The quantization processing unit 12 selects the quantization threshold THa stored in the quantization threshold storage unit 31 based on the density determination signal and the noise determination signal, and the selected quantization threshold THa and the diffusion error are added. Compare the measured pixel value. The quantization error calculated by the comparison in the quantization processing unit 12 is determined by the error determination unit 13 based on the noise determination signal, and the determination error is obtained. Finally, the diffusion coefficient setting unit 14 selects the diffusion coefficient THc from the diffusion coefficient storage unit 32 based on the determination error and the noise determination signal, and selects the diffusion error for the unprocessed next pixel as the selected diffusion coefficient THc. Calculation is performed based on the determination error, and the diffusion error is stored in the error storage unit 33.
[0053]
In the gradation reproduction processing unit 8 of FIG. 3, when the quantization processing unit 12 does not select the quantization threshold THa based on the density determination signal and the noise determination signal, that is, instead of selecting the quantization threshold THa, the error determination unit The detailed processing procedure for correcting error truncation or rounding at 13 is as follows. First, the density determination unit 34 performs density determination based on a noise determination signal for any one pixel of given image data, and outputs the density determination signal to the error determination unit 13. Next, a diffusion error for the pixel stored in the error storage unit 33 is added to the pixel value of the pixel by the adder 11 and output to the quantization processing unit 12. The quantization processing unit 12 compares a predetermined quantization threshold THa with the pixel value to which the diffusion error is added. The quantization error calculated by the comparison in the quantization processing unit 12 is determined by the error determination unit 13 based on the noise determination signal, and the determination error is obtained. Finally, the diffusion coefficient setting unit 14 selects the diffusion coefficient THc from the diffusion coefficient storage unit 32 based on the determination error and the noise determination signal, and selects the selected diffusion coefficient THc for the next unprocessed pixel and the determination The diffusion error is calculated based on the error and stored in the error storage unit 33.
[0054]
FIG. 4 is a flowchart for explaining image processing performed in the image processing apparatus 3 of FIG. In FIG. 4, among the image processing in the image processing apparatus 3 of FIG. 1, noise determination processing, noise removal processing, and gradation reproduction processing are described in detail, and descriptions of other processing are omitted.
[0055]
When encoded image data is given to the image processing apparatus 3 from the outside, the process proceeds from step S0 to step S1. In the example of FIG. 4, it is assumed that the encoded data of the image is expressed by a variable length code. First, in step S1, the variable-length code decoding unit 25 shown in FIG. 2 decodes encoded data of an image expressed by a variable-length code and separates quantized DCT coefficients. In step S2, the inverse quantization unit 26 restores the DCT coefficient by inverse quantization of the obtained quantized DCT coefficient. In step S3, the inverse orthogonal transform unit 27 performs transform to pixel data by performing inverse orthogonal transform on the restored DCT coefficient. A set of all obtained pixel data is transmitted to the input tone correction unit 41 as decoded image data. Further, only the luminance signal component Y in the decoded data of the image is given to the parameter calculation unit 22.
[0056]
Next, in step S4, the parameter calculation unit 22 calculates a noise determination parameter for determining presence / absence of noise including block noise and ringing noise using the luminance signal component Y of the decoded data of the converted image. . Note that block noise tends to occur in a flat portion of an image, and ringing noise tends to occur in an edge portion of the image. The edge part of the image refers to the periphery of the contour of the image shown in the image. For this reason, in step S5, in order to determine the presence / absence of block noise and ringing noise, the noise determination unit 7 first determines the presence / absence of noise in the flat portion and the edge portion of the image based on the noise determination parameter. If it is determined that there is noise in the flat portion of the image, then the level of block noise is obtained in step S6.
[0057]
The specific procedure of step S4 to step S6 is as follows. In the example of FIG. 4, for example, variance of pixel values in units of blocks is used as a noise determination parameter using a 3 × 3 block group as shown in FIG. The 3 × 3 block group is a 3 × 3 block group including a single target block B5 and eight peripheral blocks B1 to B4 and B6 to B9 around the target block. One arbitrary block constituting the 3 × 3 block group is composed of 64 pixels P0 to P63 for 8 rows and 8 columns, as shown in FIG. 5B, for example. The processes in steps S4 to S6 are performed in units of 3 × 3 block groups while changing the target block after the restored image is divided into one or more blocks.
[0058]
First, in step S4, based on the given luminance signal component Y, for each 3 × 3 block group shown in FIG. 5, the variances Var (1) ˜V (1) ˜ of the pixel values of all the pixels P0 to P63 in each block B1 to B9. Var (9) is calculated. The variance Var (x) of any single block Bx (x = 1 to 9) is expressed by the following equation. In the following expression, “P (i)” is a luminance signal component of the pixel value of the i-th pixel. i is an integer of 0 to 63 (i = 0 to 63), and x is an integer of 1 to 9 (x = 1 to 9).
[0059]
[Expression 1]

[0060]
Next, in step S5, first, the variance average value ALLVAR of the 3 × 3 block group to be processed is calculated. The variance average value ALLVAR of the 3 × 3 block group is an average value of variances Var (1) to Var (9) of all the blocks in the 3 × 3 block group. The variance average value ALLVAR is calculated by dividing the sum ALLSUM of the variances Var (1) to Var (9) of all blocks in the 3 × 3 block group by the number of all blocks in the 3 × 3 block group as shown in the following equation. It is a thing.
[0061]
[Expression 2]

[0062]
Next, in step S5, the variance average value ALLVar of the 3 × 3 block group to be processed is compared with the first noise determination threshold value TH1 stored in the threshold value storage unit 23 in advance. If the variance average value ALLVar is equal to or less than the first noise determination threshold value TH1, it is determined that the target block B5 of the 3 × 3 block group to be processed is located in a flat portion where noise may occur in the image, and from step S5 Proceed to step S6.
[0063]
In step S6, the variances Var (1) to Var (9) of all the blocks of the 3 × 3 block group to be processed are compared with the second noise determination threshold TH2 stored in the threshold storage unit 23 in advance. If the variances Var (1) to Var (9) of all blocks in the 3 × 3 block group to be processed are larger than the second noise determination threshold value TH2, the process proceeds from step S6 to step S10. If the variances Var (1) to Var (9) of all blocks in the 3 × 3 block group to be processed are equal to or less than the second noise determination threshold TH2, the process proceeds from step S6 to step S11.
[0064]
Specifically, the comparison process in step S6 is performed by comparing the variances Var (1) to Var (9) of each block of the 3 × 3 block group to be processed with the second noise determination threshold value TH2. That is, the variances Var (1) to Var (9) of all the blocks in the 3 × 3 block group to be processed are larger than the second noise determination threshold value TH2. All of the variances of the blocks Var (1) to Var (9) It means that it is larger than the second noise judgment threshold TH2. As described above, in step S6, the variances Var (1) to Var (9) of all the blocks are compared with the second noise determination threshold value TH2, and the level of block noise generated in the block of interest is determined. That is, a large variance Var (1) to Var (9) of all the blocks means that the level of block noise generated in the 3 × 3 block group is high.
[0065]
When the variance of all the blocks is larger than the second noise determination threshold TH2, smoothing is performed by applying spatial filter processing using the first block noise removal filter F1 to each pixel constituting the target block in step S10. Done. The first block noise removal filter F1 is a block noise removal filter having a relatively strong effect, for example, as shown in FIG. If the variance of all blocks is equal to or smaller than the second noise determination threshold TH2, smoothing is performed by applying spatial filter processing using the second block noise removal filter F2 to each pixel constituting the target block in step S11. Done. The second block noise removing filter F2 is a filter for removing block noise having a relatively weak effect as shown in FIG. 6B, for example, and has a smoothing effect at least as compared with the first block noise removing filter F1. weak.
[0066]
When the variance average value ALLVar of the 3 × 3 block group to be processed is larger than the first noise determination threshold TH1 in step S5 described above, the 3 × 3 block group to be processed at an edge portion where ringing noise may occur in the image. In order to determine whether or not the target block B5 is located, the process proceeds from step S5 to step S7. The presence / absence of ringing noise is determined in two stages in steps S7 and S8.
[0067]
In step S7, among all the blocks of the 3 × 3 block group to be processed, the variances Var (1) to Var (4), Var (6) to 8 neighboring blocks B1 to B4 and B6 to B9 excluding the target block B5 are used. It is determined whether at least one of Var (9) is less than the first noise determination threshold TH1. Furthermore, in step S7, it is determined whether or not the variance Var (5) of the block of interest is greater than or equal to the first noise determination threshold value TH1. Based on these determination results, at least one of the variances Var (1) to Var (4) and Var (6) to Var (9) of all the peripheral blocks B1 to B4 and B6 to B9 is the first noise determination threshold value TH1. If the variance Var (5) of the target block is greater than or equal to the first noise determination threshold TH1, it is determined that ringing noise has occurred in the target block B5 in the processed 3 × 3 block group, and the step Proceed from step S7 to step S12. When the variances Var (1) to Var (4) and Var (6) to Var (9) of the peripheral blocks B1 to B4 and B6 to B9 are all equal to or higher than the first noise determination threshold TH1, or the variance Var ( When 5) is less than the first noise determination threshold TH1, that is, when the condition of step S7 is not met, the process proceeds from step S7 to step S8.
[0068]
In step S8, the presence / absence of ringing noise is determined again based on conditions different from those in step S7. Specifically, using the variance Var (1) to Var (9) of all the blocks of the 3 × 3 block group to be processed and the variance VarVar of the variance value of the 3 × 3 block group to be processed, A ringing noise determination value is calculated, and the presence or absence of ringing noise is determined based on the calculated determination value. The variance VarVar of the variance value of the 3 × 3 block group is calculated by the following equation 4.
[0069]
[Equation 3]

[0070]
In step S <b> 8, the ringing noise determination value calculated based on Expression 3 is compared with the fourth noise determination threshold value TH <b> 4 stored in the threshold value storage unit 23. The fourth noise determination threshold value TH4 is 0, for example. If the determination value is larger than the fourth noise determination threshold value TH4, it is determined that ringing noise has occurred, and the process proceeds from step S8 to step S12. If the determination value is smaller than the fourth noise determination threshold TH4, it is determined that no ringing noise has occurred, and the process proceeds from step S8 to step S9.
[0071]
When the process proceeds from step S7 or step S8 to step S12, ringing noise is generated in the target block B5 of the 3 × 3 block group to be processed. Therefore, in step S12, smoothing is performed, for example, by applying a spatial filter process using a ringing noise elimination filter F3 as shown in FIG. 6C to each pixel constituting the target block B5. Is called.
[0072]
In step S9, a determination related to the smoothing process is performed on the target block B5 in which both block noise and ringing noise are not generated. In step S9, the strength of the smoothing process is adjusted based on the continuity between the blocks of the 3 × 3 block group to be processed. Specifically, when the variance Var (5) of the block of interest is equal to or less than a preset third noise determination threshold TH3, it is determined that the continuity between the blocks is relatively maintained, and the step Proceed from step S9 to step S13. In step S13, smoothing is performed by performing spatial filter processing using the first smoothing filter F4 on each pixel constituting the target block B5. The first smoothing filter F4 is a smoothing filter having a relatively weak effect as shown in FIG. 6D, for example.
[0073]
On the other hand, if it is determined that the variance Var (5) of the block of interest is larger than the third noise determination threshold TH3, it is determined that the continuity between the blocks is relatively lost, and the process proceeds from step S9 to step S14. In step S14, smoothing is performed by applying a spatial filter process using the second smoothing filter F5 to each pixel constituting the target block B5. The second smoothing filter F5 is a filter for smoothing processing having a relatively strong effect, for example, as shown in FIG. 6E, and is more effective than at least the first smoothing filter F4.
[0074]
As described above, the image processing apparatus 3 performs noise determination of block noise and ringing noise in units of 3 × 3 block groups, and performs noise removal processing based on the noise determination result. The image processing apparatus 3 performs halftone processing on the pixels in each block after performing the noise removal processing in steps S10 to S14. That is, in steps S15 to S18 in FIG. 4, error diffusion processing is performed based on the type of noise and the level of noise.
[0075]
Table 1 shows a specific example of the quantization threshold THa used in steps S15 to S18 in FIG. In the description of Step S15 to Step S18 and Table 1, it is assumed that the pixel value of the pixel is represented by 256 gradations. For example, when the pixel value is represented by 256 gradations, the low density portion pixel is a pixel having a pixel value of 0 or more and 20 or less. The pixel in the high density portion is a pixel having a pixel value of 236 or more and 255 or less. The middle density pixel is a pixel having a pixel value of 21 or more and 235 or less.
[0076]
[Table 1]

[0077]
The block of interest B5 from which block noise has been removed using the first block noise removal filter F1 in step S10 proceeds from step S10 to step S15 and is processed. In step S15, paying attention to the fact that block noise occurs at the boundary between adjacent blocks, the following error diffusion processing is performed on the pixels in the boundary two-pixel area AR1 in the target block B5. The boundary two-pixel area AR1 is an area indicated by a diagonal line rising to the right in FIG. 7, and is an area corresponding to two pixels located at the boundary portion of the target block B5.
[0078]
Specific error diffusion processing in step S15 is as follows. In the restored image, the target block B5 processed in step S10 has continuity with the adjacent blocks B1 to B4 and B6 to B9, and thus corresponds to a flat portion in the restored image. For this reason, first, in order to prevent the pixel value from being frequently inverted by the error diffusion process, the following is performed in the density determination unit 34 shown in FIG. 3 for the low density and high density pixels in the boundary 2 pixel area AR1. Such pixel value saturation processing is performed. Specifically, the pixel value is saturated to 0 for the low density pixel. Further, the pixel value is saturated to 255 for the pixels in the high density portion. The pixels in the intermediate density portion are not subjected to pixel value saturation processing, and the original pixel values are maintained. Further, instead of saturating the pixel value in the density determination unit 34, the error determination unit 13 may perform error truncation or rounding based on the density determination signal of the density determination unit 34. After the saturation process in the density determination unit 34, as described in FIG. 3, the diffusion error stored in the error storage unit 33 is added to the pixel value of each pixel of the block of interest B5.
[0079]
Next, based on the noise determination signal obtained by the noise determination performed in steps S5 to S9 and the density determination signal obtained by the density determination performed in the density determination unit 34, the quantization processing unit 12 Quantization processing is performed. Specifically, since the block of interest B5 processed in step S10 is a block in which relatively large block noise is generated, among the quantization thresholds THa stored in the quantization threshold storage unit 31, Table 1 Quantization is performed using the quantization threshold THa shown in “Error diffusion (1)”.
[0080]
In the quantization process in step S15, when the density determination unit 34 determines that the pixel to be processed is a low density part or a high density part pixel, as shown in Table 1, the level 1 quantization threshold THa is 64. And 192 is used as the level 3 quantization threshold THa. When the density determination unit 34 determines that the pixel to be processed is an intermediate density pixel, as shown in Table 1, 180 is used as the level 1 quantization threshold THa, and the level 3 quantization threshold THa. 76 is used. Thereby, in the pixels in the intermediate density portion, it is possible to easily invert the pixel values by error diffusion processing. As described above, in the quantization process in step S15, for the pixels in the intermediate density portion, the quantization threshold THa is set low on the high density side and low in order to easily cause the signal value to be inverted by the error diffusion process. Set higher on the density side.
[0081]
Next, as described above, the diffusion coefficient setting unit 14 determines the diffusion coefficient THc. The diffusion coefficient THc is determined based on the noise determination signal. For blocks in which block noise may occur, the diffusion coefficient THc of the first error diffusion range Dif1 as shown in FIG. 8A is set, and error diffusion processing is performed using the first error diffusion range Dif1. Is done.
[0082]
In the above example, the case where the pixel value of the pixel in the low density portion or the high density portion is saturated to 0 or 255 has been described. However, the error determination unit 13 may correct the quantization error to 0. . That is, the density determination unit 34 determines the density of the input pixel, and the error determination unit 13 sets the pixels determined to be the low density part or the high density part based on the obtained density determination result. The determination error may be set to 0.
[0083]
In the quantization process in step S15, the error diffusion process is performed on the pixels in the area other than the boundary 2-pixel area AR1 in the target block B5 without performing the density determination in the density determination unit 34. That is, the error diffusion process may be performed without changing the quantization threshold THa according to the density of the pixels outside the boundary two-pixel area AR1 in the target block B5.
[0084]
The block of interest B5 from which block noise has been removed using the second block noise removal filter F2 in step S11 proceeds from step S11 to step S16 and is processed. In step S16, focusing on the fact that block noise tends to remain at the boundary between adjacent blocks, the following error diffusion processing is performed on the pixels in the boundary 1 pixel area AR2 in the target block B5. The boundary 1-pixel area AR2 is an area indicated by diagonally slanting lines in FIG. 7, and is an area for one pixel located at the boundary portion of the target block B5. That is, the process of step S16 is performed on an area for one pixel located on the boundary side of the target block in the boundary two-pixel area AR1 that is the processing target in step S15.
[0085]
The specific error diffusion process in step S16 is as follows. First, the density determination unit 34 shown in FIG. 3 performs density determination in the same manner as the process performed in step S15. That is, the pixels in the low density portion or the high density portion are saturated, and the saturation process is not performed on the pixels in the intermediate density portion. Next, the quantization processing unit 12 performs quantization using the quantization threshold THa shown in “Error diffusion (2)” in Table 1. The quantization processing unit 12 performs quantization by changing the quantization threshold THa between the low density part or the high density part and the intermediate density part, similarly to the process performed in step S15. Subsequently, the diffusion coefficient setting unit 14 sets the diffusion coefficient THc of the first error diffusion range Dif1 shown in FIG. 8A based on the noise determination signal, and performs error diffusion processing using the first error diffusion range Dif1. I do. Also in step S16, as described in step S15, the density determination unit 34 does not saturate the pixel values of the low density part or high density part pixels, but the error determination unit 13 sets the quantization error to 0. You may make it set. In the quantization process in step S16, the error diffusion process is performed on the pixels in the area other than the boundary 1 pixel area AR2 in the target block B5 without performing the density determination by the density determination unit 34. That is, the error diffusion process may be performed without changing the quantization threshold THa and performing quantization according to the density of the pixels outside the boundary 1 pixel area AR2 in the target block B5.
[0086]
The block of interest B5 from which block noise has been removed using the ringing noise removal filter F3 in step S12 proceeds from step S12 to step S17 and is processed. Ringing noise is particularly likely to occur in a high-contrast portion of the pixel density regardless of the position in the block. Therefore, in step S17, pixel values are less likely to be frequently inverted due to error accumulation by error diffusion processing. The process is performed. In step S17, processing is performed on all the pixels in the block.
[0087]
The specific error diffusion process in step S17 is as follows. First, as performed in steps S15 and S16, the density determination unit 34 saturates the pixels in the low density part or the high density part, and does not perform saturation processing on the pixels in the intermediate density part. Next, the quantization processing unit 12 performs quantization using the quantization threshold THa shown in “Error diffusion (3)” in Table 1. In step S17, unlike steps S15 and S16, the same quantization threshold THa is used for all pixels in the target block B5 regardless of the low density portion, the high density portion, and the intermediate density portion. Quantization is performed. Subsequently, the diffusion coefficient setting unit 14 sets the diffusion coefficient THc of the second error diffusion range Dif2 shown in FIG. 8B based on the noise determination signal, and performs error diffusion processing using the second error diffusion range Dif2. I do. The first error diffusion range Dif1 has a wider processing range centered on the target pixel to be processed than the second error diffusion range Dif2. As described in step S15 and step S16, also in step S17, the density determination unit 34 does not saturate the pixel values of the pixels in the low density part or the high density part, but the error determination unit 13 introduces a quantization error. You may make it set to 0.
[0088]
In addition, it is determined that neither block noise nor ringing noise is generated, and the target block B5 smoothed using the first and second smoothing filters F4 and F5 in step S13 and step S14 is step S13. The process proceeds from step S14 to step S18. Since the block of interest B5 that has proceeded to step S18 is located in an area where distortion has not occurred, the density determination by the density determination unit 34 is not performed, and the quantization threshold THa shown in “Error diffusion (4)” in Table 1 is used. Thus, the quantization processing unit 12 performs quantization. That is, the quantization processing unit 12 performs the quantization using the same quantization threshold THa on all the pixels in the target block B5 regardless of the low density part, the high density part, and the intermediate density part. Subsequently, the diffusion coefficient setting unit 14 sets the diffusion coefficient THc of the second error diffusion range Dif2 shown in FIG. 8B based on the noise determination signal, and performs error diffusion processing using the second error diffusion range Dif2. I do. The process of step S18 is performed on all the pixels in the target block B5, similarly to step S17.
[0089]
As described above, in steps S15 to S18 in FIG. 4, based on the noise determination signal, the error diffusion process according to the noise type and the noise level is performed, and the density determination signal determined by the density determination unit 34 is applied. Based on the above, error diffusion processing is performed. Thus, not only the noise removal processing of the spatial filter processing unit 9 performed in steps S4 to S14 but also the error diffusion processing performed in the gradation reproduction processing unit 8, the quantization threshold THa and the error determination threshold THb or the diffusion Using a simple method of switching the coefficient THc, processing according to the type of noise and the level of noise can be performed. After error diffusion processing has been performed for all pixels in the restored image, the process proceeds from step S15 to step S18 to step S19, and the flowchart of FIG. 4 is completed.
[0090]
As described above, the image processing device 3 in FIG. 1 specifically obtains block noise by obtaining a variance for each block with respect to data obtained by decoding encoded data of an image. The presence / absence of noise typified by ringing noise, the type of noise, and the noise level are obtained. Further, the image processing apparatus 3 performs a noise removal process corresponding to each obtained noise, and further performs an appropriate halftone process corresponding to the noise. As a result, the image processing apparatus 3 can effectively remove noise from the decoded data of the image.
[0091]
Next, a more detailed configuration of the image processing apparatus 3 that executes the above-described image processing method will be described with reference to FIG. The image processing apparatus 3 in FIG. 1 includes an input tone correction unit 41, a color space conversion unit 42, a color correction unit 43, a black color in addition to the restoration processing unit 20, the gradation reproduction processing unit 8, and the spatial filter processing unit 9. A generation undercolor removal unit 44, an output tone correction unit 45, and an image storage unit 46 are included. The image forming apparatus 1 includes an operation panel 48 in addition to the image processing apparatus 3 and the image output apparatus 4. The image forming apparatus 1 including the image processing apparatus 3 and the image output apparatus 4 is realized by, for example, a so-called multifunction machine or printer. The operation panel 48 is operated by a user of the image forming apparatus 1, and the operation result is given to the image processing apparatus 3 and the image output apparatus 4 to be appropriately processed.
[0092]
Image data generated by various image input devices is input to the image processing device 3 via an interface. Examples of the given image data include encoded data of an image captured by a digital camera, and still image data captured from a moving image using software on a computer. Examples of the interface standard between the image input device and the image processing device 3 include USB (Universal Serial Bus) and IEEE 1394 (Institute of Electrical and Electronic Engineers). The encoded data of the image given to the image processing device 3 is formatted, for example, as standard data represented by JPEG.
[0093]
The image data given to the image processing device 3 is stored in the image processing device 3 in the restoration processing unit 20, the input tone correction unit 41, the spatial filter processing unit 9, the color space conversion unit 42, the color correction unit 43, and the black. The generated lower color removal unit 44, the output gradation correction unit 45, and the gradation reproduction processing unit 8 are sequentially transmitted in this order. The image data supplied to the image processing device 3 is initially a digital color signal in the form of a YCbCr signal, and finally is output from the image processing device 3 as image data in the form of a digital color signal in the form of a CMYK signal. It is output to the device 4.
[0094]
The restoration processing unit 20 decodes the encoded data of the image in the YCbCr signal format, and further performs inverse quantization processing and inverse orthogonal transform processing, thereby converting the encoded data of the image in the YCbCr signal format into the image in the YCbCr signal format. To decrypted data. Further, the restoration processing unit 20 determines whether or not block noise and ringing noise are generated in the restored image indicated by the obtained decoded data, and outputs the determination result of the noise occurrence state as a noise determination signal.
[0095]
The input tone correction unit 41 performs tone scale correction processing and color balance correction processing on the pixel data of the decoded data of the YCbCr signal system image decoded by the decoding processing unit 6 as necessary. Apply. The spatial filter processing unit 9 performs spatial filter processing using a digital filter on the image data of the YCbCr signal input from the input tone correction unit 41 based on the noise determination signal, thereby correcting the spatial frequency characteristics. Processing is performed to prevent block noise and ringing noise.
[0096]
The color space conversion unit 42 converts the color space of the decoded data of the image in the YCbCr signal format to L ^* a ^* b ^* Convert to the color space of the color system. L after color space conversion ^* a ^* b ^* The decoded data of the signal form image is input to the color correction unit 43 in the subsequent stage. The color correction unit 43 receives the given L ^* a ^* b ^* The decoded image data in the signal form is subjected to color correction in accordance with the characteristics of the image output device 4 and converted into the decoded image data in the CMY signal form after color space conversion. The decoded data of the image in the CMY signal format after color correction is input to the subsequent black generation and under color removal unit 44.
[0097]
The black generation / under color removal unit 44 performs black generation processing and under color removal processing based on the CMY signal format image data supplied from the color correction unit 43, and converts the CMY signal format image data into the CMYK signal format. Convert to image data. The black generation process is a process of generating a black signal from the three color signals of CMY included in the image data in the CMY signal format. The under color removal process is a process for generating a new CMY color signal by subtracting the black color signal from the three CMY color signals included in the image data in the CMY signal format. The image data in the CMYK signal format includes, for each pixel, a black color signal generated by the black generation process and a CMY color signal generated by the under color removal process.
[0098]
As an example of the black generation process, there is a method of generating black using skeleton black. In the method using skeleton black, black generation processing and under color removal processing are simultaneously performed on a pixel to be processed. Specifically, in the above method, the color signal of the pixel to be processed is converted as shown by the following simultaneous equations. In the following simultaneous equations, the input / output characteristic of the skeleton curve is “y = f (x)”, the input color signals of the pixels in the CMY signal form are “C”, “M”, “Y”, and the output CMYK. Each color signal of the pixel in the signal form is represented as “C #”, “M #”, “Y #”, “K #”, and the UCR (Under Color Removal) rate is represented by “α” (0 <α <1). The UCR rate α is a numerical value greater than 0 and less than 1.

[0099]
The output tone correction unit 45 performs output tone correction processing corresponding to the tone reproduction characteristics of the image output device 4 on the decoded image data in the CMYK signal format. The output tone correction process is a process for converting image data composed of a signal indicating the color density of each pixel into image data composed of a signal indicating a halftone dot area ratio, for example.
[0100]
The gradation reproduction processing unit 8 performs halftone processing which is so-called halftone generation processing on the decoded data of the image in the CMYK signal format after the output gradation correction. Further, the gradation reproduction processing unit 8 performs the above-described error diffusion processing based on the noise determination signal on the decoded data of the image in the CMYK signal format, similarly to the spatial filter processing unit 9. The halftone process is a process of determining the pixel density so that the restored image represented by the decoded data is separated into pixels, and the gradations in the image are reproduced.
[0101]
The decoded data of the image subjected to the processing in each of the processing units 20, 41, 9, 42 to 45, 8 from the restoration processing unit 20 to the gradation reproduction processing unit 8 is preferably, for example, the image storage unit 46. Are once stored, read at a predetermined timing, and input to the image output device 4. The image output device 4 visualizes the image based on the decoded data of the given image. The image output device 4 is realized by, for example, a printing device that uses a processing method represented by an electrophotographic method and an inkjet method. The printing apparatus outputs the visualized image on a recording medium realized by paper or the like. The image output device 4 is not limited to a printing device for recording media, and may be realized by an image output device 4 of another form. For example, the image output device 4 may be realized by a display device. A display device typified by a cathode ray tube (CRT) device and a liquid crystal display displays the visualized image on a screen included in the display device. In this case, in the image processing apparatus 3, the black generation and under color removal unit 44 is unnecessary and is omitted.
[0102]
FIG. 9 is a block diagram showing a configuration of an image forming apparatus 61 according to another embodiment of the present invention. FIG. 1 illustrates a configuration in which the image processing apparatus 3 includes a decoding processing unit 6, a noise determination unit 7, and a gradation reproduction processing unit 8. In the image forming apparatus 61 in FIG. 9, an image processing program 64 for causing the computer 63 to realize the functions of the decoding processing unit 6, the noise determination unit 7, and the gradation reproduction processing unit 8 is prepared in advance. The image processing apparatus 3 is realized by installing and executing the processing program 64 in a general-purpose computer 63. Thus, after decoding the encoded data of the image, the computer 63 determines the noise occurrence state in the restored image, and applies halftone processing according to the noise occurrence state to the decoded image data. As a result, the image processing program 64 can effectively remove noise generated by decoding the compression-encoded data from the decoded data using the general-purpose computer 63.
[0103]
In the image forming apparatus 61 configured as shown in FIG. 9, specifically, the computer 63 includes application software 71, a printer driver 72, a communication port driver 73, a communication port 74, and a central processing unit (CPU) 75. And a storage unit 76. In the example of FIG. 9, the image processing program 64 is divided into a program in the application software 71 and a program in the printer driver 72. The communication port 74 is realized by, for example, a communication port 74 based on the RS232C standard or a communication port 74 for LAN. The application software 71, the printer driver 72, and the communication port 74 driver 73 are stored in the storage unit 76, for example. The central processing unit 75 executes programs included in the application software 71, the printer driver 72, and the communication port 74 driver 73.
[0104]
Image encoded data represented by image data captured by a digital camera and still image data captured from a moving image is supplied to a computer 63. In the example of FIG. 9, noise determination processing and noise removal processing are executed by application software on the encoded data of a given image, and a processed restored image is output from the printer.
[0105]
The program of the application software 71 causes the computer 63 to execute the functions of the decoding processing unit 6 in FIG. 1 and the noise determination unit 7 in FIG. Preferably, the program of the application software 71 causes the computer 63 to further execute the function of the spatial filter processing unit 9 of FIG. As a result, decoding processing is performed on the encoded data of the image given to the computer 63, and noise determination processing of the restored image and a spatial filter for noise removal are performed on the obtained decoded data of the image. Processing is performed by the computer 63.
[0106]
The program of the printer driver 72 generally causes the computer 63 to execute halftone processing including noise removal processing on the decoded image data after the spatial filter processing. Specifically, the computer 63 that executes the program of the printer driver 72 performs the processing of the color correction unit 43 in FIG. 1 and the black generation and under color removal unit 44 in FIG. 1 on the image data after the spatial filter processing. The processing, the processing of the output tone correction unit 45 of FIG. 1, and the processing of the tone reproduction processing unit 8 of FIG. The computer 63 that executes the program of the printer driver 72 further realizes a printer language translation unit 77. The printer language translation unit 77 converts the decoded data of the image after halftone processing into a printer language. The decoded image data after the printer language conversion is given to the printer which is the image output device 4 via the communication port 74 driver 73 and the communication port 74.
[0107]
The image processing program 64 described above is preferably recorded on a computer-readable recording medium. If such a recording medium is used, a general-purpose computer 63 can easily use an image processing program 64 that can effectively remove noise generated by decoding of encoded image data from the decoded image data. It becomes possible to supply to. As a result, the recording medium on which the image processing program 64 is recorded can be provided in a portable manner, so that the functions realized by the image processing program 64 can be easily distributed.
[0108]
In the present embodiment, specifically, the central processing unit 75 in the computer 63 executes the image processing program 64. As a program medium that is a recording medium in which the image processing program 64 is stored, for example, a non-volatile memory used for the storage unit 76 of the computer 63 may be used. Alternatively, if the storage unit 76 of the computer 63 is provided with a medium reading device as an external storage device, it can be removably inserted into the medium reading device, and the internal program can be inserted into the medium reading device. A recording medium that can be read by the computer 63 may be used as the program medium.
[0109]
The central processing unit 75 has a configuration in which the image processing program 64 stored in the recording medium is executed by the central processing unit 75 in the computer 63 directly accessing the recording medium, for example. Alternatively, the image processing program 64 stored in the recording medium is once read from the recording medium, downloaded to the program storage area in the storage unit 76, and the downloaded image processing program 64 is sent to the central processing unit 75. It may be configured to be executed. Regardless of the two types of program media described above, the configuration of the image processing program 64 may be either of the two types described above. It is assumed that the program for downloading the program is stored in the computer 63 in advance.
[0110]
The program medium which is a recording medium storing the image processing program 64 is preferably configured to be separable from the computer 63 main body. Examples of the recording medium used as the program medium include a tape recording medium, a disk recording medium, a card recording medium, and a memory recording medium. Examples of the tape recording medium include a magnetic tape and a cassette tape. Examples of the disk recording medium include a magnetic disk medium represented by a floppy (R) disk and a hard disk device, an optical disk medium represented by a CD-ROM (Compact Disc-Read Only Memory) and a DVD (Digital Video Disc), In addition, magneto-optical disk media represented by MO (Magneto-Optic Disc) and MD (Mini Disc) can be mentioned. Examples of the card-type recording medium include an IC (Integrated Circuit) card including a so-called memory card and an optical card. Examples of the memory recording medium include a nonvolatile semiconductor memory. Nonvolatile semiconductor memories include, for example, mask ROM, EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), and flash ROM. Thus, the recording medium used for the program medium may be a medium that carries the program in a fixed manner.
[0111]
Further, the above-described image processing program 64 is not limited to the situation given to the computer 63 via a recording medium, but may be given to the computer 63 by communication via a transmission medium. Examples of the transmission medium include a transmission line of a telephone line network, a transmission line of a computer network using a LAN (Local Area Network) and a WAN (Wide Area Network), and a transmission cable for connecting computers. This makes it easier to implement the image processing device 3. Specifically, when the image processing apparatus 3 of FIG. 9 has a configuration that can be connected to a communication network including the Internet, the image processing program 64 may be downloaded from the communication network. When the image processing program 64 is downloaded from the communication network, the download program may be stored in advance in the computer 63 main body, or may be installed in advance in the computer 63 from another recording medium.
[0112]
As described above, the image processing apparatus 3 in FIGS. 1 and 9 specifically distributes the block-by-block distribution to the data obtained by decoding the compression-encoded image data. By obtaining, the presence / absence of noise represented by block noise and ringing noise, the type of noise, and the level of noise are obtained. Further, the image processing apparatus 3 performs a noise removal process corresponding to each obtained noise, and further performs an appropriate halftone process corresponding to the noise.
[0113]
The image forming apparatuses 1 and 61 and the image processing apparatus 3 according to the first and second embodiments are examples of the image forming apparatus and the image processing apparatus 3 of the present invention. It can be realized in various forms. In particular, the detailed configuration and operation of each component of the image processing apparatus 3 are not limited to the above-described configuration and operation as long as the same effect can be obtained, and may be realized by other configurations and operations. The processing unit of the image processing apparatus 3 is not limited to a configuration realized by individual circuits, but is functionally realized by causing a computer to execute control software for realizing the functions of one or more processing units. May be.
[0114]
【The invention's effect】
As described above, according to the present invention, the image processing apparatus decodes the compression-encoded image data, and determines the noise occurrence status in the image represented by the decoded data in units of blocks. Based on the result, halftone processing that not only reproduces the gradation but also reduces noise is performed on the decoded data. As a result, the image processing apparatus can effectively remove noise in the decoded data of the image while performing halftone processing.
[0115]
According to the invention, in the image processing apparatus, the image represented by the decoded data is divided into one or more blocks, and halftone processing is performed for each block. As a result, the image processing apparatus can prevent the occurrence of textures at the block boundaries in the image after halftone processing. Furthermore, according to the present invention, in the image processing apparatus, error diffusion processing is performed on the decoded data of the image as halftone processing. At least one of the three constants of the quantization threshold, the error determination threshold, and the diffusion coefficient used in the error diffusion process is switched according to the determination result of the noise occurrence state in the image. As a result, the image processing apparatus can easily prevent the occurrence of texture at the boundary of blocks in the image after halftone processing.
[0116]
According to the present invention, in the image processing apparatus, prior to the processing of the gradation reproduction processing unit, spatial filter processing for reducing noise is performed on the decoded data of the image. As a result, the image processing apparatus can more effectively remove noise from the image represented by the decoded data.
[0117]
Furthermore, as described above, according to the present invention, the image forming apparatus performs halftone processing on the decoded data of the image based on the noise generation state for each block of the image, and the halftone processing is performed. The image is visualized based on the image data. As a result, the image forming apparatus can provide a high-quality output image.
[0118]
Further, as described above, according to the present invention, the image processing method determines the decoding process of the compression-encoded image data and the noise occurrence status in the image represented by the decoded data in units of blocks. Noise determination processing is performed, and based on the determination result of the noise determination processing, halftone processing that not only reproduces gradation but also reduces noise is performed on the decoded data of the image. Thereby, when the image processing method is applied, noise generated by decoding the compression-encoded data can be effectively removed from the decoded data.
[0119]
Furthermore, as described above, according to the present invention, the image processing program decodes the function of decoding the compression-encoded data of the image and the occurrence of noise in the image represented by the decoded data in units of blocks. The computer is allowed to realize a function of determining and a function of performing halftone processing for reproducing noise and reducing noise while reproducing the gradation, on the decoded data of the image. Thus, the image processing program can effectively remove noise generated by decoding the compression-encoded data from the decoded data using a general-purpose computer. As described above, according to the present invention, the recording medium records the above-described image processing program, and the computer can read the image processing program. This facilitates the distribution of functions realized by the image processing program.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an overall configuration of an image forming apparatus 1 according to an embodiment of the present invention.
2 is a block diagram illustrating a functional configuration of a restoration processing unit 20 of the image processing apparatus 3 in FIG.
3 is a block diagram showing a functional configuration of a gradation reproduction processing unit 8 of the image processing apparatus 3 of FIG.
4 is a flowchart for explaining image processing performed in the image processing apparatus 3 of FIG. 1; FIG.
5 is a schematic diagram illustrating a configuration of a 3 × 3 block group used in noise determination processing performed in a noise determination unit 7 of the image processing apparatus in FIG. 1;
6 is a schematic diagram illustrating a configuration of a filter used in the spatial filter processing unit 9 of the image processing apparatus in FIG. 1. FIG.
7 is a schematic diagram showing a region to be processed in error diffusion processing performed in the gradation reproduction processing unit 8 of the image processing apparatus in FIG. 1; FIG.
8 is a schematic diagram showing an error diffusion range used in error diffusion processing performed in the gradation reproduction processing unit 8 of the image processing apparatus of FIG.
FIG. 9 is a block diagram showing an overall configuration of an image forming apparatus 61 according to another embodiment of the present invention.
[Explanation of symbols]
1,61 Image forming apparatus
3 Image processing device
4 Image output device
6 Decryption processing unit
7 Noise judgment part
8 gradation reproduction processing section
9 Spatial filter processing section
11 Adder
12 Quantization processing section
13 Error judgment part
14 Diffusion coefficient setting part
63 computer
64 Image processing program
THa Quantization threshold
THb Error judgment threshold
THc diffusion coefficient
B5 attention block
B1-B4, B6-B9 peripheral block

Claims (8)

A compression processing unit that receives compression-encoded data representing an image and decodes the compression-encoded data;
The image represented by the decoded data is divided into a target block including a plurality of pixels and peripheral blocks around the target block, and a variance of pixel values of all pixels is calculated for each block, and threshold processing is performed based on the calculated variance. Whether or not noise is generated in the block is determined for each block, and for the block in which noise is generated, the type of noise and the degree of occurrence of the noise are determined , and based on the determination result A noise determination unit for selecting a noise removal filter to be used for spatial filter processing ;
A gradation reproduction processing unit that performs halftone processing for reducing the noise while reproducing the gradation based on the determination result of the noise determination unit, on the decoded data of the image, An image processing apparatus.   The gradation reproduction processing unit divides an image represented by the decoded data into one or more blocks including a plurality of pixels, and performs the halftone processing for each block. Image processing device. In order for the gradation reproduction processing unit to perform halftone processing by performing error diffusion processing on the decoded data of the image,
(A) based on the decoded data of the image, for each pixel constituting the image, an adder that obtains a diffusion pixel value that is the sum of a pixel value and a diffusion error;
(B) By quantizing the diffusion pixel value of each pixel based on a predetermined quantization threshold, a quantization value to be the pixel value of the halftone processed image data is obtained for each pixel, and diffusion is performed A quantization processing unit that further obtains a quantization error, which is a difference between the pixel value and the quantization value, for each pixel;
(C) an error determination unit that corrects a quantization error based on a predetermined error determination threshold;
(D) a diffusion coefficient setting unit that calculates a diffusion error for allocating a corrected quantization error to an unprocessed pixel value based on a predetermined diffusion coefficient;
The image processing apparatus according to claim 1, wherein at least one of the quantization threshold, the error determination threshold, and the diffusion coefficient is switched based on a determination result of the noise determination unit. A spatial filter processing unit that performs spatial filtering on the decoded data of the image based on the determination result of the noise determination unit;
The image processing apparatus according to claim 1, wherein the gradation reproduction processing unit performs halftone processing on image data that has been subjected to spatial filter processing. An image processing device according to any one of claims 1 to 4,
An image forming apparatus, comprising: an image output device that receives image data subjected to halftone processing and visualizes the image based on the image data. A process of decoding compression-encoded data representing an image;
The image represented by the decoded data is divided into a target block including a plurality of pixels and peripheral blocks around the target block, and a variance of pixel values of all pixels is calculated for each block, and threshold processing is performed based on the calculated variance. Whether or not noise is generated in the block is determined for each block, and for the block in which noise is generated, the type of noise and the degree of occurrence of the noise are determined , and based on the determination result A process of selecting a noise removal filter to be used for the spatial filter process ;
An image processing method comprising: performing halftone processing for reducing noise while reproducing gradation based on a determination result of noise on the decoded data of the image. A function of decoding compression-coded data representing an image;
The image represented by the decoded data is divided into a target block including a plurality of pixels and peripheral blocks around the target block, and a variance of pixel values of all pixels is calculated for each block, and threshold processing is performed based on the calculated variance. Whether or not noise is generated in the block is determined for each block, and for the block in which noise is generated, the type of noise and the degree of occurrence of the noise are determined , and based on the determination result A function of selecting a noise removal filter used for spatial filter processing ;
An image characterized by causing a computer to realize a function of performing halftone processing for reducing noise while reproducing gradation based on a determination result of noise on the decoded data of the image Processing program.   A computer-readable recording medium on which the image processing program according to claim 7 is recorded.

Images