201003537 九、發明說明： 【發明所屬之技術領域】 本發明係有關一種拍照方法，特別是一種臉部辨識 之連續拍照方法。 【先前技術】 目前，數位相機使用者普遍運用按壓快門方式拍照， 而無論是拍攝他人或自拍，按壓快門之拍照方式皆會造成 手震情況發生，因此目前相機多附有防手震功能。 其中，光學式防手震功能是在鏡頭系統中加入浮動之 感光元件或浮動之鏡片，使鏡頭保持穩定而不會因晃動產 生模糊影像。電子式防手震功能則是藉由電子電路修正因 晃動產生的模糊影像，達到影像修正的效果。 然而，無論是光學式或電子式防手震功能皆無法完全 避免手震的影響，因此便發展出透過辨識被拍攝者之臉部 微笑而自動拍照的技術，其能避免按壓快門產生手震而影 I 響影像品質，且能立即補捉到被拍攝者的微笑表情。· 辨識臉部微笑的自動拍照技術是一種透過數位相機 内建韌體的運算與辨識技術，其鎖定被拍攝者之臉部呈現 微笑特徵的三角位置，並於偵測到臉部之微笑特徵時自動 進行拍照。 然而，目前數位相機之辨識臉部微笑的自動拍照技 術僅限定於偵測微笑表情，對於微笑以外之其它表情則 並無支援，若被拍攝者表情不斷變化（例如：吐舌或皺眉 201003537 等）則相機將無法連續地 對微笑表情進行拍攝，因二情，僅能針 【發明内容】 w了數位相機的應用。 為了解决上述先前技術之問題，本發明接ά褲人 部辨識之連續拍照方法，以解=^出一縣 __表情致使相機無法連先^技= 所存在僅限定 外的其它表情的問題。去連，，地自動拍攝微笑以 ，以供種臉部辨識之連續拍照方法，用 攝者的臉部輪;;Γ反t包括透過輪 量二:，之臉部影像以得到影像變化 其中，土斟本 艾動拍祕鏡項拍攝被拍攝者。 之臉部影像 化量超過拍攝啟叙f 士 聶乂知則包括當影像變 〇 此外：啟動拍攝鏡頭拍攝被拍攝者。 再者，更二::::物攝被拍攝者之次數。 -人數计异完成時停止取得臉部影像。之人放田 另外，還可加入步驟包括當影 動值=:於二_停止取得於拍攝啟 偵測被拍攝者的臉之連續拍照方法藉由 部_臉部影像，再㈣相:兩::::== 201003537 =影像變化量，最後再㈣影像變化量啟動拍攝鏡頭 被拍攝者，因此只要被拍攝者的臉部表情不斷變 者=鏡頭將自動進行連續拍照，且得以拍攝到被拍 躡者的各種表情。 /以上之關於本發明内容之說明及以下之實施 明係用以示範盥解鏗才八 ° 明夕直—、㈣本發明之精神與原理，並且提供本發 月之寻利申請範圍更進一步之解釋。 【實施方式】 點，科㈣本發明之詳細特徵以及優 内容熟習相關技藝者了解本發明之技術 w 以’且根據本說明書所揭露之内容、申社專 相關之目的及優點=輕易地理解本發明 明之以任何觀點限制本發細爾發 #芩肊「第丨圖」，為本發明— 之連續拍照流程圖。如「第】圖所“例之月以辨識 續拍照方法包括先透過拍攝」不部辨識之連 廓（步驟100)，並及考抱〜兄、測破拍攝者的臉部輪 驟110)，接著i设侍具有臉部輪廓的臉部影像（步 影像變化量次取得之臉部影像以得到 攝鏡頭拍攝被拍攝者（步驟叫根據衫像變化量啟動拍 中的器官以===，=兩次取得之臉部影像 里。拍攝步騍則包括當影像變 201003537 化量=拍攝 :動值時啟動拍攝鏡頭拍攝被拍攝者。 u例為透過拍攝鏡頭偵測被 並反覆取得被拍攝者之臉部影像，再根撐比較相， 得之臉部影像中的器官所得c兩次取 Φ，^ 如1冢艾化1大於拍摄啟 值Ψ啟動拍攝鏡頭拍攝被拍攝者。 心以下舉例說明，例如，當拍攝者欲拍攝被拍攝者時， 後攝者之臉部輪-步驟_，缺 俊反後取得被拍攝者之臉部影像（步驟 摄“'、 臉部表情變化時，其臉部影像也將隨i 7賴者之 :取二之臉部嶋的器官之位置或角度將會出：差 二二者t由比對相鄰兩次取得之臉部影像中的器官便 侍1衫像變化量（步驟120；)，最後杏& f & 過拍攝啟動值β 士則” 仔之影像變化量超 驟 =立即欲動拍攝鏡頭拍攝被拍攝者（步 像。便可連續地拍攝到被拍攝者的各種表情影 凊參照「第2 F! 之連續拍照流程圖明二施例之臉部辨識 部辨識之白叙所不，本實施例之臉 的臉部輪穿κ步包t透㈣攝鏡頭偵測被拍攝者 (步驟200彳 0)，牧者設定拍攝被拍攝者之次數 (牛= ，再來便反覆取得具有臉部輪廓的臉部影像 鏡頌拍攝被_者=之後’根據影像變化量敌動抽攝 聶者（步驟130)，最後，計算拍攝被拍攝者 201003537 之次數，當次數計算完成時 210)。 7 得臉部影像（步驟 本實施例為透過拍攝鎊 接著設定拍攝被拍攝者之次教、’被拍攝者之臉部輪廓， 部影像，跟著根據比較相鄰兩覆取得被拍攝者之臉 影像變化量啟動拍攝鏡 :侍之臉部影像所得之 攝被拍攝者之次H ㈣拍攝者。最後，計算拍 像。 “數叶算完成時停止取得臉部影 以下舉例說明，例如，去 將先透過拍攝鏡頭伯、、丨 田 々奴拍攝被拍攝者時， 後、'，、彳拍攝者之臉部輪廟（步驟100)，铁 後叹疋拍攝被拍攝者之 Uj ^ 攝之目的，之㈣p )，错以達成連續拍 110)，當被拍攝4入^得被拍攝者之臉部影像（步驟 之改變，所以相；二部表情變化時’其臉部影像也將隨 接著h μ 得之臉部影像將會出現差盈， 化量相鄰兩絲得之臉部影像便得到影像變 拍躲據影像變化量啟動拍攝鏡頭拍攝被 數達到先=，錢計算拍攝被拍攝者之次數，當次 臉部影像拍攝被拍攝者之次數時即停止取得 的各種表情影像:’ 3此，便可連㈣拍攝職拍攝者 之連圖」，為本發明—實施例之臉部辨識 部辨識之自_日1 弟圖」所不，本實施例之臉 …、方去包括先透過拍攝鏡頭偵測被拍攝 201003537 者的臉部輪逾彳此 罕两甸（步驟]00)，並 臉部影像（步騍】〗 仅取传具有臉部輪廓的 部影偉以彡e f 者再比對相鄰兩次取得之月& 丨心像以传到影像變化量(步驟 取仔之化 化量小於拍攝啟動值之次數 )—E者，當影像變 部影像(步驟300),曰％ 、、頜疋值時停止取得臉 頭拍攝被拍攝者（步驟]3())。h像文化里啟動拍攝鏡 本實施例為透過拍攝鏡頭土 並反覆取得被拍摄 、、川被拍鉍4之臉部輪廓， r 辑者之月我部影像，再比齡相卿 「臉部影像而得到冉比較相相次取得之 攝啟動值之讀然後當影像變化量小於拍 像，最後，再根據夺 了止1，貝取传臉部影 者。 /〜又化里啟動拍攝鏡頭拍攝被拍攝 以下舉例說明，例如，各 將先透過拍攝鏡頭心… U欲拍攝被拍攝者時， 後反承致、、川被拍攝者之臉部輪廓（步驟100),铁 俊反覆取得被拍攝者之 7…、 脸部士卜主^ 〇如像（步驟110)，當被拍攝者之 卩衣6受化時，其臉部影 ± ^ 乜氷丨逍之改支，所以相鄰兩201003537 IX. Description of the invention: [Technical field to which the invention pertains] The present invention relates to a photographing method, and more particularly to a continuous photographing method for face recognition. [Prior Art] At present, digital camera users generally use the shutter to take pictures, and whether shooting a person or self-timer, pressing the shutter to take a picture will cause a jitter situation, so the camera is often equipped with anti-shake function. Among them, the optical anti-shake function is to add a floating photosensitive element or a floating lens to the lens system to keep the lens stable without causing blurred images due to shaking. The electronic anti-shake function is to correct the blurred image caused by the shaking by the electronic circuit to achieve the effect of image correction. However, neither the optical or electronic anti-shake function can completely avoid the effects of jitter, so a technique of automatically taking a picture by recognizing the smile of the face of the subject is developed, which avoids the occurrence of jitter by pressing the shutter. Shadow I sounds in image quality and instantly captures the smiling expression of the photographer. · The automatic camera technology for recognizing facial smiles is a computing and recognition technology built into the firmware of the digital camera, which locks the triangle position of the face of the photographer with a smile feature, and detects the smile feature of the face. Take pictures automatically. However, the current automatic camera technology for recognizing facial smiles of digital cameras is limited to detecting smiling expressions, and there is no support for expressions other than smiles, if the subject's expression changes constantly (for example: tongue or frown 201003537, etc.) Then the camera will not be able to continuously shoot the smiling expression, because of the two emotions, only the needle [invention content] w digital camera application. In order to solve the above problems of the prior art, the continuous photographing method of the human body of the trousers of the present invention solves the problem that the camera cannot be connected to the first __ expression. Going to connect, the ground automatically shoots a smile, for the continuous photographing method for facial recognition, using the photographer's face wheel;; Γ 反 includes the transmission of the image of the face through the wheel 2: The bandit Benjamin shot secret shot was taken by the photographer. The amount of facial imaging is higher than that of shooting. Nie Zhizhi includes when the image changes. 〇 In addition: Start the shooting lens to shoot the subject. Furthermore, the second:::: The number of times the object was taken by the photographer. - The face image is stopped when the number of people is different. In addition, the person can also add steps, including when the shadow value =: on the second _ stop to obtain the continuous photographing method for detecting the face of the subject by shooting, by the part _ face image, and then (four) phase: two: :::== 201003537=Image change amount, and finally (4) Image change amount starts shooting the subject, so as long as the subject's facial expression changes continuously = the lens will automatically take continuous photos, and can be photographed The various expressions of the leader. The above description of the present invention and the following embodiments are intended to demonstrate the spirit and principles of the present invention and to provide further scope for the application for the benefit of the present month. Explanation. [Embodiment] Points, Sections (4) Detailed Features and Advantages of the Invention The skilled artisan understands the technology of the present invention, and understands the contents and advantages of the disclosure according to the contents disclosed in the present specification. The invention expressly limits the present invention to the continuous photographing flow chart of the present invention. For example, in the example of "the month of the example, the method of identifying the continued photographing method includes first photographing through the photographing" (step 100), and the method of holding the brother, measuring the face of the photographer 110), Then i sets the face image with the contour of the face (the image of the face image obtained by the step change is obtained to obtain the camera for the camera (the step is to start the organ in the shot according to the amount of change of the shirt image ===,= In the facial image obtained twice, the shooting step includes starting the shooting lens to shoot the subject when the image is changed to 201003537. The shooting is detected by the shooting lens and the photographer is repeatedly obtained. The facial image, then the roots are compared, and the organ obtained in the facial image is taken twice. Φ, ^ If 1冢 Aihua 1 is larger than the shooting threshold, the shooting lens is taken to shoot the subject. For example, when the photographer wants to shoot the subject, the back of the photographer's face wheel - step _, the face image of the subject is obtained after the lack of the camera (step "", when the facial expression changes, the face The image will also follow the i 7: take the face of the second The position or angle of the official will be: the difference between the two is determined by comparing the amount of the organ in the facial image obtained by the two adjacent faces (step 120;), and finally the apricot & f & Start value β 士 ” 影像 之 之 影像 影像 = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = In the second embodiment of the flowchart, the face recognition part of the embodiment is not described. The face of the face of the embodiment is rotated through the κ step package to detect the subject (step 200彳0), the shepherd. Set the number of times the person is photographed (Niu =, and then repeatedly obtain a face image with a facial contour, and the subject is photographed by _ = after 'the enemy' according to the amount of image change (step 130), and finally, Calculate the number of times the photographer is photographed 201003537, when the number of times is calculated 210). 7 Get the face image (Steps in this example are to shoot the pound and then set the subject's secondary teaching, 'the subject's face outline, Part of the image, followed by comparing the two adjacent shots to get taken The face image change amount starts the shooting mirror: the image taken by the waiter's face image is taken by the photographer H (four) photographer. Finally, the image is calculated. "When the number of leaves is calculated, the face image is stopped. Going to shoot the photographer through the shooting lens, and 丨田々奴, then, ', 彳 彳 彳 ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( ( Purpose, (4) p), wrong to achieve continuous shooting 110), when being photographed 4 into the face image of the photographer (the step changes, so the phase; when the two expressions change), the facial image will follow h μ get the facial image will appear poor, the amount of the adjacent two facial images will get the image change to capture the image change amount to start the shooting lens to be counted first =, money calculation shot by the photographer The number of times, when the number of times the face image is captured by the photographer, the various expression images that are stopped are obtained: '3, then the picture of the photographer can be connected (4), which is the face recognition of the present invention. The identification of the Ministry from the _ day 1 brother map The face of this embodiment, the party includes the first shot through the shooting lens to detect the face of the 201003537 person's face wheel over this rare two steps (steps 00), and the face image (step 〗) only take the pass The silhouette of the facial contour is compared with the & 再 再 再 再 再 再 & & & & & & & & & & & & & & & & & & & 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻 相邻When the image is changed (step 300), 曰%, and jaw value are stopped, and the face is photographed (step 3). h In the culture, the shooting mirror is activated. This embodiment is to capture the lens and capture the contours of the face, and the image of the face of the Sichuan 被 , , , , , , 我 我 我 我 我 我 我 我 我 我 我 我 我 我 我 我 我 我And the 启动 冉 冉 取得 取得 取得 取得 当 当 当 当 当 当 当 当 当 当 当 当 当 当 当 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像 影像For example, each of them will first pass through the lens of the shooting lens. When the U wants to shoot the subject, the face of the photographer is reversed (step 100), and Tie Jun repeatedly obtains the subject of the photographer. ..., the face of the master, like the image (step 110), when the subject's coat 6 is affected, the face shadow ± ^ 乜 丨逍 改 change, so the adjacent two
Sr之臉部影像將會出現差異，接著藉由比對相鄰： 之臉部影像便得到影像變化量 量小於拍攝啟動值之次數大於額定值時^ 、、"… 心驟300)’以避免數位相機資源之 浪費’取後再根據當所得 攝被拍攝者（步驟130)。 一、交化夏啟動拍攝鏡頭拍 以下將舉例以臉部辨識之連續拍照裝置對本發明之方 10 201003537 法做進一步說明。請參照 例之臉部辨識之連續拍日⑽w _ 1」，為根據本發明—實施 臉部辨識之連續拍照裝置;=圖。如「第4圖」所示， 幕30。 1已括機身10、拍攝鏡頭20與螢 本貫施例為透過拍攝 廓，並反覆取得被拍 …侦測被拍攝者之臉部輪 次取得之臉部影像所π ”部影像’再根據比較相鄰兩 拍攝被拍攝者。㈣化量啟動拍攝鏡頭20 以下舉例說明，例如，杏 將先透過拍攝鏡頭2G貞:&拍攝被拍攝者時， 並將被婦之—顯步驟⑽） 拍攝者之次數（步驟細），藉 、^疋拍攝被 後透過拍攝鏡頭20反覆取得被為之目的，然 _。當被㈣者之臉_ 步驟 =广相鄰兩次取得之臉部影二== :位i或角度將會出現差異。其中 :：之相對 例如肌肉變化產生之敏紋官的差異 以及嘴巴由閉合至張開等。接著^由目目月半閉或全閉 得之臉部影像中的兩次取 影像變化量小於拍攝啟動值之次數大於额當 繼縯取得臉部影像（步驟3〇〇) 、、疋％停止 費。當所得之影像變化量超過拍資源之浪 拍攝鏡頭2◦拍攝被拍攝者(步驟13〇)。最料算== 201003537 攝者之次數，當次數達到先 * 數時即停止取得臉部影像（步^=拍攝被拍攝者之次 拍攝到被拍攝者的各種表情影像7 )。因此，便可連續地 於實際應用上，本發 :、 現，因此使用者可將二:=目:暢體或軟體實 拍照模式。當使用者 '痛式切換成連續自動 者時，將先透過拍攝鏡^自動拍照模式拍攝一被拍攝 ,拍攝者之臉部輪廓顯^於^皮拍攝^之臉部輪靡並將被 r次數（例如：三次）’接著:π、，跟者設定拍攝被拍攝者之 臉部影像與第二臉部影像臉部輪廉的第— 時，其臉部影像也將隨之改變，因此=:_化 影像變化量啟動拍攝鏡頭拍攝被 為一次。爯#，雜磘^ 卫·汁α拍知次數 旦1 '取得被拍攝者之臉部輪廓 影像，且比對第二臉部影像與第^Sr's facial image will be different, and then by comparing the adjacent: the facial image, the image change amount is less than the shooting start value is greater than the rated value ^, , "... heart 300) Avoiding the wastage of digital camera resources' is taken after the camera is taken (step 130). First, the cross-border summer shooting lens shot The following will be exemplified by the face recognition continuous photographing device to further explain the method of the present invention 10 201003537. Please refer to the continuous shooting day (10)w _ 1" of the face recognition of the example, which is a continuous photographing device for performing face recognition according to the present invention; As shown in Figure 4, Curtain 30. 1 Included in the body 10, the shooting lens 20 and the firefly embodiment are through the shooting profile, and repeatedly obtain the captured image... detect the face image obtained by the photographer's face round π "section image" and then according to Compare the two adjacent photographers. (4) Start the shooting lens 20 The following example shows, for example, apricot will first shoot through the shooting lens 2G贞:& when shooting the subject, and will be photographed by the woman-display step (10) The number of times (steps are fine), borrowing, ^疋 shooting is repeatedly obtained through the shooting lens 20 for the purpose, then _. When the face of (four) is _ step = wide adjacent to get the face shadow two = = : There will be a difference in position i or angle. Among them: the difference between the sensitive striatum caused by muscle changes and the mouth from open to open, etc. Then the face is half-closed or fully closed. The number of changes in the image taken twice in the image is less than the number of times the shooting start value is greater than the amount of time to obtain the face image (step 3〇〇), 疋% stop fee. When the resulting image change exceeds the wave of shooting resources 2◦ Shoot the subject (step 13〇). Count == 201003537 The number of photographers, when the number reaches the first * number, the face image is stopped (step ^ = shooting the photographer to capture the various facial expression images of the photographer 7). Therefore, it can be continuously In practical applications, this is:, now, so the user can put two: = eye: smooth or soft real camera mode. When the user 'painfully switch to continuous auto, will automatically take pictures through the lens ^ When the mode is shot, the face of the photographer's face is displayed on the face of the camera and will be r times (for example: three times). Then: π, the follower sets the face of the photographer. When the image and the second facial image face are inferior, the facial image will also change accordingly, so the =: _ image change amount is activated and the shooting lens is taken once. 爯#,杂磘^ 卫· The juice α beats the number of times 1 ' to obtain the silhouette image of the face of the subject, and compares the second face image with the second
I 像變化量，再根據第二影像 :二影 攝者，且計算拍攝次數為二次。最後^ 拍 之臉部輪廊的第四臉部影像，且㈣取得被拍攝者 臉部影像而得到第三影像L:比=與第四 :動拍:鏡頭拍攝被拍攝者，且計算拍 止繼續取得臉1旦彡俨入（幻如··二久），因此便停 像。其中，臉部影像由複數個像素單元 組成，因此可時出兩個臉部影像之間像素單元Γ差異 12 201003537 置’因而得到影像變化量 檔案格式，可 (BMP; Bit Mapped) 點之門旦次4/ 乂兩個BMP檔案之影像中每一個圖 活占之，日1影像貧料的差 ^ 例性說明，_ 、而侍知影像變化量，然上述僅為示 “非用以限定本發明之實施態樣。 :亦可同時偵測多人之臉 ._之•然其僅為 明之實施態樣。 非用以限疋本發 r ,綜上所述，本發明之臉部辨識之連續拍日〃方+基士 '谓測被拍攝者的臉部輪廓，反覆取得错 部輪广•的臉部影像，再比對相鄰兩次取 传到影像變化量，最後再根據影像變化量啟動拍 拍攝被拍攝者，因此 里欠助拍攝鏡頭 化，拍摄# α ώ ^要被拍扣者的臉部表情不斷變 拍祕鏡碩將自動進行連續拍照--攝者的各種表情。 于拍祕到被拍 #然本發_前述之實_揭露如上H (限定本發明。在不脱離本發明之精神和範圍戶 =用以 動與濁飾，均屬本發明之專利保 …之更 定之保護範圍請參考所附之申請專利範圍本發明所界 【圖式簡單說明】 乐1圖係為本發明—實施例之臉 圖，· κ〈是，拍照流程 乐2圖係為本發明—實施例之臉 圖； 两艾連績拍照流程 201003537 第3圖係為本發明一實施例之臉部辨識之連續拍照流程 圖；以及 第4圖係為本發明一實施例之臉部辨識之連續拍照裝置示 意圖。 【主要元件符號說明】 10 ...........................機身 20 ...........................拍攝鏡頭 30 ...........................螢幕 14The I image is changed according to the second image: the second image is taken, and the number of times of shooting is calculated twice. Finally ^ shoot the fourth face image of the face wheel gallery, and (4) obtain the face image of the subject to obtain the third image L: ratio = and fourth: motion capture: the lens captures the photographer, and calculates the beat Continue to get the face once into the (illusion · · two long), so it stops. The face image is composed of a plurality of pixel units, so that the pixel unit Γ difference between the two face images can be set to 12 201003537, thus obtaining the image change amount file format, which can be (BMP; Bit Mapped) The second 4/ 乂 two BMP files in the image of each of the maps live, the difference between the image and the image of the day 1 example, _, and the amount of image change, but the above is only for the purpose of "not used to limit this The embodiment of the invention can also detect the faces of a plurality of people at the same time. However, it is only for the implementation of the invention. It is not limited to the present invention. In summary, the face recognition of the present invention is The continuous shooting of the Japanese side + Kishi's is to measure the face contour of the photographer, and repeatedly obtain the face image of the wrong part of the wheel, and then compare the two times to the image change amount, and finally according to the image change The amount is activated to shoot the subject, so it is not helpful to shoot the lens, shooting #α ώ ^The face expression of the person to be beaten is constantly changing. The mirror will automatically take continuous pictures--the various expressions of the photographer. Shoot the secret to be photographed #然本发_ The above facts _ expose as above H (limit the invention Without departing from the spirit and scope of the present invention, the use of the invention and the illuminating of the invention are all the scope of protection of the patents of the present invention. Please refer to the attached patent application. Le 1 is the face diagram of the invention - the embodiment, κ < is, the photographing process music 2 is the face diagram of the invention - the embodiment; the two Ai continuous performance photographing process 201003537 the third figure is a invention The continuous photographing flowchart of the face recognition of the embodiment; and the fourth figure is a schematic diagram of the continuous photographing device for the face recognition according to an embodiment of the invention. [Description of main component symbols] 10 .......... .................body 20 ...........................shooting lens 30 ...........................screen 14