TW201227608A - Method and system for rendering multi-view image - Google Patents
Method and system for rendering multi-view image Download PDFInfo
- Publication number
- TW201227608A TW201227608A TW99145933A TW99145933A TW201227608A TW 201227608 A TW201227608 A TW 201227608A TW 99145933 A TW99145933 A TW 99145933A TW 99145933 A TW99145933 A TW 99145933A TW 201227608 A TW201227608 A TW 201227608A
- Authority
- TW
- Taiwan
- Prior art keywords
- image
- view
- program
- new
- view image
- Prior art date
Links
Landscapes
- Image Processing (AREA)
Abstract
Description
201227608 TW6973PA· 六、發明說明: 【發明所屬之技術領域】 本發明疋有關於-種影像之建構方法及系統,且特別 是有關於-種多視角影像之建構方法及系統。 【先前技術】 —數位影像具有「不浪費底片、不佔空間、永不褪色、 谷易儲存、方便攜帶、容易編修」等優點,使得數位影像 • 已經逐漸取代傳統底片所拍攝之照片。201227608 TW6973PA· VI. DESCRIPTION OF THE INVENTION: 1. Field of the Invention The present invention relates to a method and system for constructing an image, and more particularly to a method and system for constructing a multi-view image. [Prior Art] - Digital image has the advantages of "no waste of negative film, no space, no fading, easy storage, easy to carry, easy to edit", so that digital images have gradually replaced the photos taken by traditional negatives.
Ik著數位影像科技的發展,各種影像祕技術也不斷 的發展。透過影像編修技術可以美化照片,或者加入有趣 的圖案,甚至是編修成一多視角立體景多像。 然而,多視角立體影像的建構方法相當的複雜。以處 理技術而言,需要有效地提升處理速度,使得多視角立體 影像更能普遍被接受。 籲【發明内容】 根據本發明之一方面,提出一種多視角影像之建構方 法。多視角影像之建構方法包括以下步驟。提供數張單视 角影像,此些單視角影像之視角不同。對此些單視角景J 進f 一解析度調整程序(Resizing),以獲得數個新解析 =影像之至少部份像素。對此些新解析度影像之此至少告 伤像素進行一交錯排列程序(View Interlace),以开< 士 一張多視角影像。 夕 201227608Ik is developing digital image technology, and various video secret technologies are constantly developing. Through image editing technology, you can beautify your photos, or add interesting patterns, or even edit them into a multi-view stereoscopic image. However, the construction method of multi-view stereoscopic images is quite complicated. In terms of processing technology, it is necessary to effectively increase the processing speed, so that multi-view stereoscopic images are more generally accepted. SUMMARY OF THE INVENTION According to one aspect of the present invention, a method of constructing a multi-view image is proposed. The method for constructing a multi-view image includes the following steps. Several single-view images are provided, and the viewing angles of these single-view images are different. For these single-view scenes, a resolution adjustment program (Resizing) is performed to obtain a plurality of new resolutions = at least some pixels of the image. For this new resolution image, at least the pixel is scanned for a View Interlace to open a multi-view image. Evening 201227608
TW6973PA 根據本發明之另一方面,提出—種,-系統。多視角影像之建構系統包括一提供m衫像之建構 元。提供單元用以提供數張單視角影像:處理單 之視角不同。處理單元係對此些單視 :士現角影像 調整程序(Resizing),、 行一解析度 部份像f㈣L 新解析度影像之至少TW6973PA According to another aspect of the invention, a system is proposed. The multi-view image construction system includes a construction element that provides a m-shirt image. The unit is provided to provide several single-view images: the processing angle is different. The processing unit is a single view of this: the Resizing image, the resolution of the image, and the resolution of the F (4) L new resolution image.
抽像素,並對此麵解析度影像 ^之至J 交錯排列程序(View Inte細)素進行-像。 形成一張多視角影 為讓本發明之上述内容能更明顯易懂 實施例,並配合賴圖式,作詳細說明如下了文特舉各種 【實施方式】 以下係提出實施例進行詳細說明, 範例說明,並不會限縮本發明欲保護之範圍。^以作為 術特點。 走顯不本發明之技 第一實施例 請參照第1圖’其、㈣本案第—實_之 簡之建構系統_的示意圖(多視角影像piQ4 第3圖)。多視角影像建構系統1〇〇包括一提供單 ^ 及了處理單元12〇。提供單元11G用以提供各種資訊, 如疋一照相機、-攝影機…儲存雜或連接儲存媒體 一連,埠。處理單元120用以執行各種處理、運算程序 例如疋中央處理器晶片(Central Processing Unit 201227608 TW15973PA* CPU) GPU) 、— 圖形處理器晶片(Graphic Processing Unit, 韋刃體電路或儲存數組程式碼之一儲存媒體。本實 施例之處料兀12G包括數組第1行緒ΐ2ι及數组第二 執^緒122,第-執行緒121係以平行處理之 種處理、運算程序’第二執行緒⑵也是以平行處理之方 式進行各種處理、轉㈣。 請參照第2〜3圖,第2圖繪示第—實施例之多視角 ,04之建構方法之流程圖’第3圖繪示第_實施例 之夕視角影像P1G4之建構過程的示意圖。以下更搭配一 ^程圖及-示意圖說明本實施例之多視角。 及本實施例之多視角影像_之 法。然而,本案所屬技術領域中具有通常知識者均 多視角影像P1°4之建構系統_並不 P1 , 〇4之建構方法亦不侷限應㈣帛 P104之建構系統1〇〇。 ’夕現角心像Draw a pixel and perform an image-like image on the face resolution image to the J staggered program (View Inte). The present invention is described in detail with reference to the drawings, and The description does not limit the scope of the invention to be protected. ^ as a feature of surgery. The present invention is directed to the first embodiment of the present invention. The multi-view image construction system 1 includes a providing unit and a processing unit 12A. The providing unit 11G is used to provide various information such as a camera, a camera, a storage device, or a connection storage medium. The processing unit 120 is configured to execute various processing and computing programs, such as a central processing unit (Central Processing Unit 201227608 TW15973PA* CPU) GPU, and one of a graphics processing unit (Graphic Processing Unit, or a stored array code) The storage medium. In the embodiment, the material 12G includes an array of the first line ΐ 2 ι and the second array of the array 122, the first thread 121 is processed in parallel processing, and the operation program 'the second thread (2) is also In the parallel processing, various processes and transitions are performed (4). Please refer to FIGS. 2 to 3, FIG. 2 is a flow chart showing the multi-view of the first embodiment, and the construction method of 04. FIG. 3 is a diagram showing the first embodiment. A schematic diagram of the construction process of the illuminating view image P1G4. The following is a more detailed description of the multi-view of the embodiment and the multi-view image of the embodiment. However, the general knowledge in the technical field of the present invention is The construction system of multi-view image P1°4 _ is not P1, and the construction method of 〇4 is not limited to (4) 建P104 construction system 1〇〇.
Pioi ^^Γρι〇+ιPioi ^^Γρι〇+ι
=如R」代表紅色、「G」代表綠色、「B」代表藍色) 其解析度為640x360x3。 J 接著,在步驟S102中,處理單元12〇之第一執行 121依據深度資訊,以平行處理之方式對原始影像 行像素平移㈣(Pixel Rendering) A1A像素插補輕= (HoleFilling) A2,以形成數張新視角影像^”。 μ第 201227608= If R" stands for red, "G" stands for green, and "B" stands for blue). The resolution is 640x360x3. J, then, in step S102, the first execution 121 of the processing unit 12 performs Pixel Rendering A1A pixel interpolation light (HoleFilling) A2 in parallel processing according to the depth information to form Several new perspective images ^". μ第201227608
TW6973PA 3圖為例’原始影像舰經過像素平移程序M及像素插 補程序A2後’形成8張新視角影像P1G2。其中每-張新 2影像P102之職也是64G,行數也是湖色域通道 疋3 (「R」代表紅色、「G」代表綠色、%代表藍色), 其解析度為640x360x3。 其中,像素平移程序^用以將原始影像?1〇1的每一 =平移至適當的位置,以糊建構出8張新視角影像 02。在像素平移過程令,新視角影像⑽可能會出現 缺口的現象’此時則透過像素插補程序A2來填補缺口。 以第3圖為例,單視角影像⑽包括t張原始影像 P101及8騎視角料簡,所以單視心彡像㈣總共 ^張,這些單㈣影像⑽之視角皆不相同,但解析 度^為640x360x3。所以’透過步驟_〜siThe TW6973PA 3 picture is taken as an example. After the original image ship passes through the pixel shifting program M and the pixel interpolation program A2, eight new viewing angle images P1G2 are formed. Each of the new 2 images P102 is also 64G, and the number of lines is also the lake color gamut channel 疋3 ("R" stands for red, "G" stands for green, % stands for blue), and its resolution is 640x360x3. Among them, the pixel shift program ^ used to the original image? Each of 1〇1 is translated to the appropriate position to construct 8 new viewing angle images 02. In the pixel shifting process, a new view image (10) may have a gap phenomenon. At this time, the gap is filled by the pixel interpolation program A2. Taking Figure 3 as an example, the single-view image (10) includes t original images P101 and 8 riding angles, so the single-view image (4) is a total of ^, the viewing angles of these single (four) images (10) are different, but the resolution ^ is 640x360x3. So 'through the steps _~si
了數張單視角影像P110。 T 然後’在步驟S103中,此些第一執行緒121以平行 ^理之方式對此些單視角影像_進行一解析度調整程 序(Resizing) A3,以獲得數個新解析度影像ρι〇3。以第 3圖為例,9張單視㈣像P11G的解析度皆為_χ36㈣, 而母一張新解析度影像P103之列數為192〇,行數為 1080,色域通道為3,其解析度為192〇χΐ〇8〇χ3。 在本實施例中,像素平移程序A1、像素插補程序Μ 及解析度調整程序A3皆利用第一執行緒121以平行處理 之方式執行。以第3圖為例,原始影像ρι〇1與新視 像P102之間的箭頭表示各個第—執行緒m執行像素^ 移程序A1及像素插補程序A2之步驟sl〇2的動作,所以 201227608 TW6973PA. *此Ϊ新Si Ϊ I21之數目係為新視角影像P1 02之列數 八此二新視角景,像p 1 〇 2之張| i本&、 之張數的乘積(即640x8)。 再者,新視角影像P102與新 箭頭表示各個第-執行緒121 ^解析度衫像P103之間的 步_ 3的動作,所以此,第執―行執解_析:,調整程广 夂梱虹、日& —弟執仃緒121之數目也為 乘積^1=2之列數與新視_piG2之張數的 頭表干另卜像_與新解析度影像P103之間的箭 ::理早兀’例如是中央處理器 ^Several single-view images P110 are shown. Then, in step S103, the first threads 121 perform a resolution adjustment procedure (Resizing) A3 on the single-view images _ in a parallel manner to obtain a plurality of new resolution images ρι〇3. . Taking Figure 3 as an example, the resolution of nine single-view (four) images like P11G is _χ36 (four), and the number of columns of the new resolution image P103 is 192 〇, the number of rows is 1080, and the gamut channel is 3. The resolution is 192 〇χΐ〇 8 〇χ 3. In the present embodiment, the pixel shifting program A1, the pixel interpolation program Μ, and the resolution adjustment program A3 are all executed in parallel by the first thread 121. Taking FIG. 3 as an example, the arrow between the original image ρι〇1 and the new video P102 indicates the operation of the steps s1〇2 of the pixel-shifting program A1 and the pixel interpolation program A2 for each of the first-execution m, so 201227608 TW6973PA. *The number of new Si Ϊ I21 is the number of new perspective images P1 02, the new perspective, like the product of p 1 〇2 | i &, the number of sheets (ie 640x8) . Furthermore, the new view image P102 and the new arrow indicate the action of the step _ 3 between the respective first-execution 121 and the resolution-type shirt P103, so this is the first step of the execution of the action, and the adjustment of Cheng Guanghong, The number of the day & the younger brother 121 is also the arrow between the number of products ^1=2 and the head of the new _piG2, and the arrow between the new resolution image P103: Early 兀 'for example, the central processing unit ^
Pr〇ceSSlngUnit,cpu)來同步執行。 接^在步驟S104中’第二執行緒122以平行處理 (Vi=析度影像⑽之像素進行-交錯排列程序 第3_W^terlaCe) A4’以形成—張多視角影像P104。以 圖為例,新解析度影像P1〇3 之解析度相同(=。=)成 選出對;*執彳了緒122只需直接在崎析度影像P1G3中挑 之:==即可組成多視角影像。ι™ 如第3圖所示,新解析度影像Pl〇3與多視角影僮ριη/1 ^ ^頭表示交錯排列程序A4之步驟S1G4的動作。所 多相Γ㈣122之數目係為多視角影像P104之列數、 χ3)。角影像PH)4之行數及色域通道的乘積(即192〇_ 如第3圖所示,多視角影像ρι〇4所組成之像素皆執 201227608Pr〇ceSSlngUnit, cpu) to synchronize execution. In step S104, the second thread 122 is processed in parallel (Vi = pixel of the resolution image (10) - staggered program 3_W^terlaCe) A4' to form a multi-view image P104. Taking the figure as an example, the resolution of the new resolution image P1〇3 is the same (=.=) to select the pair; *Exercising the thread 122 only needs to pick directly in the resolution image P1G3: == can form a lot Perspective image. ιTM As shown in Fig. 3, the new resolution image P10 and the multi-view movie ριη/1 ^ ^ head indicate the operation of the step S1G4 of the interleaving program A4. The number of multiphases (four) 122 is the number of columns of the multi-view image P104, χ 3). The angle image PH) is the product of the number of lines and the color gamut channel (ie 192 〇 _ as shown in Figure 3, the pixels composed of the multi-view image ρι〇4 are executed 201227608
TW6973PA :::巧調整程序,3,所以多視角 ’ 择^由單視角影像削3之數個鄰 之每-像 ::因此多視角影像pl〇4之每一像素二點運算所 現出t位置應該呈現的内容。 、^夠精準地表 叫參照第4圖,其繪示另一實施例 之建構過程的示意圖。以第4 影像 =像素平移轉(PlxelRendeHm ’2料_ 序(Holejr彳η. N 1及像素摘補鞋 原始影像P10〗卿)形成數張新視角影像P102,後, 4P101及新視角影像P10 後’ 列程序(Vi 丹直接進行交錯排 hew Interlace) A4 來獲得 」〇4 ’而沒有經過解析度調整程序(見角办像 於多視角影像⑽,所組叙像^^ΚΓ1Ζΐη§)Α3。由 程序A3,多# P1ru,像素未執仃過解析度調整 多視角影像蘭之每一像素 P11〇之像素直接挑選。因此多視角 早視角/像 素都較無法精準地表現出此位置上 =4之每-像 來,晝面上將呈現鑛齒現象。㈣呈現的内容。如此一 皆經由反S3旦圖二施二多視角影像_之每一像素 得,戶ΐΓί 個鄰近像㈣浮點運算所獲 出此視角影像麗之每一像素都能夠精準地表現 該呈現的内容,進而使得畫面上的雜齒現象可 第 施例 實施例之多視角 請參照第5〜7圖,第5圖繪示第二 201227608 TW6973PA· 影像P104之建構系統的示意圖,第 施例之多視角影像顧之建構方法之流程圖、,,第=實 不第二實施例之多视角影像P1〇4之建構過程的干土 = % 本:施例之多視角影像P104之建構系統2〇。及方 一實施例之多視角影像_之建構系统議盘方法^第 =於執行緒之執行方式的安排,其餘相同之處不再同重TW6973PA ::: Qiao adjustment program, 3, so multi-view 'select ^ from a single-view image to cut 3 of each of the neighbors - like:: therefore, multi-view image pl 〇 4 each pixel two-point operation is shown t The content that the location should present. , which is precisely described with reference to Figure 4, which shows a schematic diagram of the construction process of another embodiment. The fourth image = pixel shift (PlxelRendeHm '2 material _ order (Holejr彳η. N 1 and pixel picking shoes original image P10〗) to form a number of new perspective images P102, after 4P101 and new perspective image P10 'The column program (Vi Dan directly interleaved hew Interlace) A4 to get "〇4' without the resolution adjustment procedure (see the corner image like multi-view image (10), the group image ^^ΚΓ1Ζΐη§) Α3. Program A3, more # P1ru, the pixel has not been subjected to the resolution adjustment multi-view image blue, each pixel P11 〇 pixel directly selected. Therefore, the multi-view early view / pixel can not accurately represent this position = 4 For each image, the mineral tooth phenomenon will appear on the surface. (4) The content presented. So through the anti-S3 Dan Figure 2, the two multi-view image _ each pixel, the household ΐΓ 邻近 邻近 neighboring image (four) floating point arithmetic Each pixel that obtains this view image can accurately represent the content of the presentation, and thus the phenomenon of the tooth on the screen can be referred to the fifth to seventh figures in the embodiment of the embodiment. FIG. Second 201227608 TW6973PA· Shadow Schematic diagram of the construction system of P104, the flow chart of the construction method of the multi-view image of the first embodiment, and the dry soil of the construction process of the multi-view image P1〇4 of the second embodiment: % For example, the multi-view image P104 construction system 2〇 and the embodiment of the multi-view image _ the construction system discussion method ^ the second arranging the execution mode of the thread, the rest of the same place is no longer the same
如第5圖所示’本實施例之處理單元22〇包括 一執行緒221、數個第二執行緒222及數個/第 223。第-執行緒221用以執行像素平移程序 Rendering) A1 及像素插補程序(H〇le Fiuing) μ ^第二執行緒222用以執行解析度調整程序(㈣動 3之動作,第三執行緒223用以執行交錯排列程序^ Interlace) Α4 之動作。 w ,以下係搭配第6及第7圖詳細說明本實施例之多 影像P104之建構方法。在步驟㈣中,提供單元u 供原始影像P101及原始影像P1〇1之深度資訊。 接著,在步驟S202中,處理單元22〇之第一執行緒 221依據深度資訊,以平行處理之方式對原始影像^❾丨進 行像素平移程序Μ及像素插補㈣A2,以形成數張新視 角影像P102。 在本實施射,像素平移㈣A1及像素姉程序A2 皆利用第-執行緒221以平行處理之方式執行1第7圖 為例’原始影像P101與新視角影像P102之間的箭頭表示 各個第-執行緒221執行像素平移程序A1及像素插補程 201227608As shown in Fig. 5, the processing unit 22 of the present embodiment includes a thread 221, a plurality of second threads 222, and a plurality of / 223th. The first thread 221 is used to execute the pixel shifting program Rendering) A1 and the pixel interpolation program (H〇le Fiuing) μ ^ The second thread 222 is used to execute the resolution adjustment program ((4) motion 3, third thread 223 is used to perform the action of the interlace program ^ Interlace) Α 4. w, the following is a detailed description of the construction method of the multi-image P104 of the present embodiment in conjunction with the sixth and seventh figures. In step (4), the unit u is provided with depth information for the original image P101 and the original image P1〇1. Next, in step S202, the first thread 221 of the processing unit 22 performs pixel shifting process and pixel interpolation (4) A2 on the original image in parallel processing according to the depth information to form a plurality of new viewing angle images. P102. In the present embodiment, the pixel shifting (4) A1 and the pixel 姊 program A2 are all performed by the parallel processing in the first thread 221 by using the first thread 221 as an example. The arrow between the original image P101 and the new view image P102 indicates each first execution. 221 executes pixel shifting program A1 and pixel interpolation process 201227608
TW6973PA 序A2之步驟S202的動作,所以此些第一執 I童 目係為新視角影像P102之列數與此' 張數的乘積(即64〇x8)。 〜新現角影像簡之 然後,在步驟S203中,此些第二執杆 A處3理視角影像,進行解析度調:程: A3以獲侍數個新解析度影像ρι〇3。 緒22f 2施Γ中,解析度調整程序A3係利用第二執行 、,者222以平行處理之方式執行。以第?圖 =象=原始影像Ρ101與物 之間的箭頭表示各個第二執行 = 序A3的動作’所以此些第二執行_之 影像ρ、=@Λ料^像則3之驗、各㈣解析度 之張數Π 域通道數及此些新解析度影像_ 數的乘積(即 1920x1080x3x9)。 接著,在步驟S204中,第三執行緒2以 之方式對新解析度影像削3之傻音推^23以千饤處理 Α4,以形Μ、θ Γ 订交錯排列程序 ^ ^〇3 4,ΓΛ?.ΡΙ04°""7 度相同,因此®^ 》成之纽㈣像P1G4之解杆 P103中挑1屮^灯緒223只需直接在新解析度影像 如、ί應位置之像素即可組成多視角影像pi〇4。 ,箭頭 緒223之叙F听以第二執行 ⑽之行^ #視角影像⑽之列數、多視角影像 仃數及色域通道數的乘積(即192〇χ1〇8〇χ3)。 201227608The action of step S202 of the TW6973PA sequence A2 is such that the first number of children is the product of the number of columns of the new view image P102 and the number of the 'sheets (i.e., 64 〇 x 8). - The new angle image is simplified. Then, in step S203, the second handle A is used to analyze the angle of view, and the resolution is adjusted: A3 to obtain a number of new resolution images ρι〇3. In the case of 22f 2, the resolution adjustment program A3 is executed by the second execution using the second execution. To the first? Figure = image = original image Ρ 101 and the arrow between the object indicates the action of each second execution = sequence A3 'so the image of the second execution _, ρ, =@Λ^^, 3, and (4) resolution The number of sheets is the product of the number of domain channels and the number of these new resolution images (ie 1920x1080x3x9). Next, in step S204, the third thread 2 pushes the silly sound of the new resolution image 3 by 饤4, and arranges the interlaced program ^^〇3 4 by the shape θ, θ ,, ΓΛ?.ΡΙ04°""7 degrees are the same, so ®^ 》成之纽(四) like P1G4's solution rod P103 pick 1 屮 ^ lamp 223 only need to directly in the new resolution image such as, ί should position the pixel A multi-view image pi〇4 can be formed. , arrow 223, F, listen to the second execution (10) trip ^ #View image (10) number of columns, multi-view image number of products and the number of gamut channels (ie 192 〇χ 1 〇 8 〇χ 3). 201227608
TW6973PA 第三實施例 請參照第8〜1〇圖,第8圖繪示第三實施例之多視角 影像P104之建構系統300的示意圖,第9圖繪示第三實 施例之多視角影像P104之建構方法之流程圖,第1〇圖繪 示第三實施例之多視角影像P1〇4之建構過程的示意圖。 本實施例之多視角影像P104之建構系統300及方法與第 一實施例之多視角影像P1〇4之建構系統1〇〇與方法不同 之處在於執行緒之執行方式的安排,其餘相同之處不再重 • 複敘述。 如第8圖所示’本實施例之處理單元32〇包括數個第 一執行緒321及數個第二執行緒322。第一執行緒321用 以執行像素平移程序(Pixel Rendering) A1及像素插補 程序(Hole Filling) A2之動作,第二執行緒322用以執 行解析度調整程序(Resizing) A3及交錯排列程序(View Interlace) A4 之動作。 暑 以下係搭配第9及10圖詳細說明本實施例之多視角 影像P104之建構方法。在步驟S301中,提供單元11〇提 供原始影像P101及原始影像P101之深度資訊。 接著,在步驟S302中,處理單元32〇之第一執行緒 321依據深度資訊,以平行處理之方式對原始影像ρι〇ι進 行像素平移程序(Pixel Rendering) “及像素插補程序 (Hole Filling) A2,以形成新視角影像ρι〇2。 在本實施例中’像素平移程序A1及像素插補程序A2 皆利用第一執行緒321以平行處理之方式執行。以第1〇 圖為例,原始影像P101與新視角影像Pl02之間的箭頭表 11 201227608TW6973PA, please refer to FIG. 8 to FIG. 1 for the third embodiment, FIG. 8 is a schematic diagram of the construction system 300 of the multi-view image P104 of the third embodiment, and FIG. 9 is a multi-view image P104 of the third embodiment. A flowchart of the construction method, and FIG. 1 is a schematic diagram showing a construction process of the multi-view image P1〇4 of the third embodiment. The construction system 300 and method of the multi-view image P104 of the present embodiment is different from the construction system 1 and method of the multi-view image P1〇4 of the first embodiment in the arrangement of the execution mode of the thread, and the rest are the same. No longer repeat • Retelling. As shown in Fig. 8, the processing unit 32 of the present embodiment includes a plurality of first threads 321 and a plurality of second threads 322. The first thread 321 is configured to perform a Pixel Rendering A1 and a Hole Filling A2 action, and the second thread 322 is configured to perform a Resizing A3 and a staggered program ( View Interlace) The action of A4. The following is a detailed description of the construction method of the multi-view image P104 of the present embodiment in conjunction with FIGS. 9 and 10. In step S301, the providing unit 11 provides depth information of the original image P101 and the original image P101. Next, in step S302, the first thread 321 of the processing unit 32 performs Pixel Rendering on the original image ρι〇 in parallel processing according to the depth information. “And the pixel filling program (Hole Filling) A2, to form a new view image ρι〇2. In the embodiment, the 'pixel shifting program A1 and the pixel interpolation program A2 are all executed in parallel by the first thread 321 . Taking the first image as an example, the original Arrow between image P101 and new perspective image P1022 Table 11 201227608
TW6973PA 第-執行謂執行像素平移程序Ai及像素插補 之步驟S302的動作’所以此些第一執行緒%】之 數2為新視角影像P102之列數與此些新視角影像謂 之張數的乘積(即640x8)。 然後,在步驟S綱t,此些第二執行緒322以平行 之方式對此些單視角影像ρί1〇進行解析度調整程序 、(ResiZing) A3及交錯排列程序⑽^獻)a4, 以直接獲得多解析度影像PI04。 在本實施例中,第二執行緒322僅針對單視角影像 # P110之部分像素進行解析度調整程序A3,而沒有實際完 成任何-張新解析度影像觸,故第1G圖以虛線標示虛 擬的新解析度影像p1〇3。其中虛擬的新解析度影像P103 並不完整,只有部份的像素被完成,這些像素被完成後, 則直接進行交錯排列程序A4安插至多視角影像ρι〇4内。 單視角影像P110(包含原始影像P101與新視角影像P102) 與多解析度景>像?1 〇4之間的箭頭表示各個第二執行緒322 執行解析度調整程序A3及交錯排列之步驟S3〇3的動作,· 所以此些第二執行緒322之數目係為多視角影像ρι〇4之 列數、多視角影像P104之行數及色域通道數的乘積(即 1920x1080x3)〇 上述實施例之單視角影像P110係由原始影像ρι〇1經 由像素平移程序(Pixel Rendering)八丨及像素插補程序 (Hole Filling) A2來獲得新視角影像P1〇2後所組成, 然而在其他實施例中,單視角影像p11〇係可以直接由數 個不同視角的鏡頭所拍攝,而不需要進行像素平移程序A1 12 201227608 TW6973PA* 及像素插補程序A2。 以上實施例係有關於一種多視角影像之建構方法及 系統,其利用平行處理之方式來提升多視角影像之處理速 度,並利用解析度調整程序來提升多視角影像之品質。 綜上所述,雖然本發明已以各種實施例揭露如上,然 其並非用以限定本發明。本發明所屬技術領域中具有通常 知識者,在不脫離本發明之精神和範圍内,當可作各種之 更動與潤飾。因此,本發明之保護範圍當視後附之申請專 • 利範圍所界定者為準。 【圖式簡單說明】 第1圖繪示本案第一實施例之多視角影像之建構系 統的示意圖。 第2圖繪示第一實施例之多視角影像之建構方法之 流程圖。 第3圖繪示第一實施例之多視角影像之建構過程的 m 不意圖。 第4圖繪示另一實施例之多視角影像之建構過程的 示意圖。 第5圖繪示第二實施例之多視角影像之建構系統的 不意圖。 第6圖繪示第二實施例之多視角影像之建構方法之 流程圖。 第7圖繪示第二實施例之多視角影像之建構過程的 示意圖。 13 201227608TW6973PA The first execution executor performs the pixel shifting procedure Ai and the pixel interpolation step S302. The number 2 of the first thread % is the number of columns of the new view image P102 and the number of the new view images. The product of 640x8. Then, in step S, the second threads 322 perform a resolution adjustment procedure, a (ResiZing) A3, and a staggered program (10) (a) for the single-view images ρί1〇 in parallel to obtain directly Multi-resolution image PI04. In this embodiment, the second thread 322 performs the resolution adjustment program A3 only for a part of the pixels of the single-view image # P110, and does not actually complete any new resolution image touch, so the 1G map is indicated by a dotted line. The new resolution image p1〇3. The virtual new resolution image P103 is not complete, only some of the pixels are completed. After these pixels are completed, the interlaced program A4 is directly inserted into the multi-view image ρι〇4. Single-view image P110 (including original image P101 and new view image P102) and multi-resolution scene > image? The arrows between 1 and 4 indicate that the respective second threads 322 perform the operations of the resolution adjustment program A3 and the staggered arrangement S3〇3, so that the number of the second threads 322 is the multi-view image ρι〇4 The number of rows, the number of rows of the multi-view image P104, and the number of gamut channels (ie, 1920x1080x3). The single-view image P110 of the above embodiment is composed of the original image ρι〇1 via the Pixel Rendering and the pixel. Hole Filling A2 is used to obtain the new view image P1〇2. However, in other embodiments, the single view image p11 can be directly taken by several different angles of view without the need for pixels. Translation program A1 12 201227608 TW6973PA* and pixel interpolation program A2. The above embodiment relates to a method and system for constructing a multi-view image, which uses parallel processing to improve the processing speed of multi-view images, and uses a resolution adjustment program to improve the quality of multi-view images. In view of the above, the present invention has been disclosed in various embodiments, and is not intended to limit the present invention. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention. Therefore, the scope of the invention is defined by the scope of the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic view showing the construction system of the multi-view image of the first embodiment of the present invention. Fig. 2 is a flow chart showing the construction method of the multi-view image of the first embodiment. Fig. 3 is a view showing the construction process of the multi-view image of the first embodiment. FIG. 4 is a schematic diagram showing the construction process of the multi-view image of another embodiment. Fig. 5 is a view showing the construction of the multi-view image construction system of the second embodiment. Fig. 6 is a flow chart showing the construction method of the multi-view image of the second embodiment. Fig. 7 is a view showing the construction process of the multi-view image of the second embodiment. 13 201227608
TW6973PA 示意圖。 第8圖繪示第二實施例之多視角影像之 ϋ。 攝系統的 方法之 第9圖繪示第三實施例之多視角影像之建構 流程圖。 第10圖繪示第三實施例之多視角影像 示意圖。 過程的 【主要元件符號說明】 100、200、300 :多視角影像之建構系統 110 :提供單元 120、 220、320 :處理單元 121、 22卜321 :第一執行緒 122、 222、322 :第二執行緒 223 :第三執行緒 Α1 .像素平移程序(pixei Rendering) A2 ·像素插補程序(jj〇ie pilling) A3 :解析度調整程序(Resizing) Μ父錯排列程序(View Interlace) plol、Pl〇i’ :原始影像 P102、P102’ :新視角影像 P103:新解析度影像 P1()4 ' P104’ :多視角影像 P110 .單视角影像 S1〇1〜Sl〇4、S201 〜S204、S301 〜S303 :流程步驟Schematic diagram of the TW6973PA. Figure 8 is a diagram showing the multi-view image of the second embodiment. Fig. 9 is a flow chart showing the construction of the multi-view image of the third embodiment. Fig. 10 is a view showing a multi-view image of the third embodiment. [Main component symbol description] 100, 200, 300: multi-view image construction system 110: providing unit 120, 220, 320: processing unit 121, 22 321 : first thread 122, 222, 322: second Thread 223: Third Thread Α 1. Pixel Rendering A2 • Pixel Interpolator (jj〇ie pilling) A3: Resizing View Repling plol, Pl 〇i' : original image P102, P102': new view image P103: new resolution image P1() 4 ' P104' : multi-view image P110. single-view image S1〇1 to S1〇4, S201 to S204, S301~ S303: Process steps
Claims (1)
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW99145933A TW201227608A (en) | 2010-12-24 | 2010-12-24 | Method and system for rendering multi-view image |
CN2011100357316A CN102567975A (en) | 2010-12-24 | 2011-02-01 | Construction method and system of multi-view image |
US13/110,105 US20110216065A1 (en) | 2009-12-31 | 2011-05-18 | Method and System for Rendering Multi-View Image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW99145933A TW201227608A (en) | 2010-12-24 | 2010-12-24 | Method and system for rendering multi-view image |
Publications (1)
Publication Number | Publication Date |
---|---|
TW201227608A true TW201227608A (en) | 2012-07-01 |
Family
ID=46413326
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW99145933A TW201227608A (en) | 2009-12-31 | 2010-12-24 | Method and system for rendering multi-view image |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN102567975A (en) |
TW (1) | TW201227608A (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106331670A (en) * | 2015-06-30 | 2017-01-11 | 明士股份有限公司 | Endoscope stereoscopic visualization system and method by employing chromaticity forming method |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CA2380105A1 (en) * | 2002-04-09 | 2003-10-09 | Nicholas Routhier | Process and system for encoding and playback of stereoscopic video sequences |
EP1437898A1 (en) * | 2002-12-30 | 2004-07-14 | Koninklijke Philips Electronics N.V. | Video filtering for stereo images |
TW200937344A (en) * | 2008-02-20 | 2009-09-01 | Ind Tech Res Inst | Parallel processing method for synthesizing an image with multi-view images |
CN101610424B (en) * | 2009-07-13 | 2011-12-21 | 清华大学 | Method and device for synthesizing stereo image |
-
2010
- 2010-12-24 TW TW99145933A patent/TW201227608A/en unknown
-
2011
- 2011-02-01 CN CN2011100357316A patent/CN102567975A/en active Pending
Also Published As
Publication number | Publication date |
---|---|
CN102567975A (en) | 2012-07-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5011316B2 (en) | Rendering the output image | |
CN100565589C (en) | The apparatus and method that are used for depth perception | |
US20140192171A1 (en) | Method for displaying three-dimensional integral images using mask and time division multiplexing | |
JP2008537250A (en) | Depth detection apparatus and method | |
US20120170833A1 (en) | Multi-view image generating method and apparatus | |
CN1765133A (en) | Three-dimensional television system and method for providing three-dimensional television | |
CN1745589A (en) | Video filtering for stereo images | |
CN1942902A (en) | Ghost artifact reduction for rendering 2.5D graphics | |
JP2011523743A (en) | Video signal with depth information | |
US20110157311A1 (en) | Method and System for Rendering Multi-View Image | |
WO2013082832A1 (en) | Method and device for image processing | |
KR20200051599A (en) | Image processing device and file generation device | |
CN106937103B (en) | A kind of image processing method and device | |
KR100989435B1 (en) | Method and apparatus for processing multi-viewpoint image | |
US10602120B2 (en) | Method and apparatus for transmitting image data, and method and apparatus for generating 3D image | |
TW201227608A (en) | Method and system for rendering multi-view image | |
CN110910485B (en) | Immersion type cave image manufacturing method | |
JP2013520923A (en) | Split screen for 3D | |
CN103473754A (en) | Novel mural digitalized archiving method | |
US20110216065A1 (en) | Method and System for Rendering Multi-View Image | |
US10551632B2 (en) | Panel device and display device | |
US8374463B1 (en) | Method for partitioning a digital image using two or more defined regions | |
KR20210094646A (en) | Immersive video bitstream processing | |
KR101794492B1 (en) | System for displaying multiview image | |
KR101801898B1 (en) | Method and apparatus for generating representing image from multi-view image |