1.353792 « v 九、發明說明： 【發明所屬之技術領域】 本發明係關於一種視訊編碼的方法，尤其是一種降低 移動估測（Motion Estimation)的運算量之視訊編碼的方 法。 【先前技術】 ® 數位影像已經是非常普遍用來傳播資訊的媒介，為 了能更有效率地節省影像處理時間以及儲存空間，往往 就會使用影像壓縮技術。 影像壓縮技術主要是利用移動估測（Motion Estimation)運算，此一運算是希望找出影片中物體移動的 軌跡，以期在後續的移動補償（Motion Compensation)能以 最少的資料量重建影像（frame)。 鲁 更進一步而言，視訊資料是由許多影像（frame)所組 合成的，當我們在看影片的時候，那些連續動作其實都 是由一張張影像（frame)所構成的，利用我們的視覺暫 留，讓我們以為那些動作是連續的。因此這一張張的影 . 像中，相鄰的兩張影像的大部分區域都相當相像，利用 這個特性，將兩張影像的每個像素（pixel)相減，得到一個 差值（difference)，而因為兩張影像相似度高，所以相減 後的值也會很小，如果我們只存第一張影像，第二張就 僅儲存差值，如此一來需要儲存的資料量就會減少很 5 13537921.353792 « v IX. Description of the Invention: [Technical Field of the Invention] The present invention relates to a method of video encoding, and more particularly to a method of reducing video encoding of a calculation amount of Motion Estimation. [Prior Art] ® Digital imagery is already a very popular medium for disseminating information. In order to save image processing time and storage space more efficiently, image compression technology is often used. Image compression technology mainly uses the Motion Estimation operation, which is to find the trajectory of the object moving in the film, in order to reconstruct the image with the minimum amount of data in the subsequent Motion Compensation. . In addition, video data is composed of many frames. When we watch a movie, those continuous actions are actually composed of a single frame, using our vision. For the time being, let us think that those actions are continuous. Therefore, in this image, most of the adjacent two images are quite similar. Using this feature, each pixel (pixel) of the two images is subtracted to obtain a difference. Because the two images have high similarity, the subtracted value will be small. If we only save the first image, the second one will only store the difference, so the amount of data that needs to be stored will be reduced. Very 5 1353792
多。而再還原的時候只要將前一張影像減掉差值就可p 得到後一張的值了。 ’ U 而連續影片中一定會有移動的物體，也就是相 個物體在前後兩張的位置不同。如果能找到物體移_ # 位置，再做相減’則得到的差值就會越小。因此就還^义 須找到移動向量（Motion vector)，而估測物體移動到哪裡 的方法就是移動估測（Motion Estimation )。還原影像時， 依據移動向量找到要相減差值的影像區塊就稱為移動# 償(Motion Compensation) ° 移動向量簡單來說就是估算目前影像（Current Frame)裡 ΝχΝ(例如 8x8 或 16x16)的區塊（Macro Block) 與前一張影像（Previous Frame)裡最相似的參考區塊 (Reference Block)位移的向量。將目前影像裡要用來估算 的Macro Block和Reference Block裡相同大小的某― 區塊作相減，然後將相減後的每個像素取絕對值相加得 到一個值，稱為絕對差值和（SAD)，其中SAD值越小就 代表這兩個區塊越相近。將搜尋區域中所有的位置都算 出一個SAD值，最小的值即是所要尋找的移動後的位 置，如此便可算出移動向量（Motion Vector)。 雖然，利用移動向量與差值可減少需儲存的運算資 料量，然而比對兩張影像找出移動向量仍需要不少的運 算量。另外，由於影像壓縮的技術架構為熟習本項技術 領域者所周知，因此未在此多做贅述。 6 1.353792 【發明内容】 本發明之主要目的係在於提供一種視訊編碼的方 法。 本發明之另—目的係在於提供一種降低影像廢縮技 術中移動估測（motion estimation)的運算量之視訊編碣的 為達到上述之目的，本發明提供一種視訊編碼的方 法。視訊由複數個影像（frames)所組成， 含複數個區塊，該視訊卜物體具有—位置㈣^ 法包括以下步驟： 步驟S11H取該物體分财第—f彡像與第二影像之位置 資訊。例如位置資訊可以是來自一作業系統（0 s)。 步驟S12:依據該物體之位置資訊，計算含有該物體之區 。較佳者’是將該第二影像中含有該 U二區塊定義為畫面間區塊（inter_block)，並依據 該預估移動向量以進行編碼。在本實施例中，本發明可 ^據》亥物體之該些位置資訊，計算該物體之位置或尺寸 诗^進纟將因為該物體位置或尺寸之變化而對應在 該第=影像中新出現的區塊定義為晝面内區塊 (intra-block)，以進行編碼。 步驟⑴：依據該預估移動向量，對該視訊進行編碼。 =體在第-影像之位置f訊包括—識別碼，且該 5 疋用以辨識物體。因此藉由作業系統所提供的識 7 1353792 別，’即可辨識物體’ u 了解是要擷取哪一個物體的位 置資訊，如此即可節省運算量。 ‘該物體具有一視窗框，利用作業系統的位置資訊可 獲知該物體在該第一影像之四個座標值，也就是該視窗 框之四個轉角（corner)位置的座標值。如眾所周知者作 業系統可提供視窗資訊，因此關於擷取座標值的技 多贅述。 物體從第一影像移動到第二影像有多種可能實施 例，包括物體單純的移動及物體縮放（scale)。因此，若是 物體單純的移動，則物體的尺寸在該第一影像與該第二 影像實質上相同。若是物體縮放，則物體的尺寸^第二 影像與第二影像不同，且分別為放大或縮小物體的尺寸。 在一實施例中，該視訊為一作業系統晝面之操作過 程，且物體之位置資訊是由該作業系統提供。 、 除了上述物體單純的移動及物體縮放之外，還可新 增或刪除物體。因此本發明還揭露在新增物體時，降低 移動估測的運算量之視訊編碼的方法，該視訊由複數個 影像（frames)所組成，每一該些影像包含複數個區塊，該 視訊中一物體具有一新增物體資訊及一位置資訊，其^ 法包含： 步驟S21 ··擷取物體之新增物體資訊。較佳者，新増物體 資訊是來自一作業系統（0S)。 _ 步驟S22:當該物體於一第一影像中不存在，而新增於一 8 1353792 第二影像時，擷取該物體在該第二影像中之該位置資訊。 步驟S23:將該影像含有該物體之區塊定義為畫面内區 塊，以對該視訊進行編碼。 另外，本發明亦揭露在刪除物體時，降低移動估測 的運算量之視訊編碼的方法，其方法包含： 步驟S31 :絲物體之—刪除物體f訊。較佳 體資訊是來自一作業系統。 〗除物 步驟S32:當該物體於—第—影像中存在，而在二男 像中被刪除時，擷取該物體在第一影像中之位置資訊。^ 步驟S33:將該第二影傻巾斜座 現的區地m金像巾對應物體刪除後所新出 的α “義為旦面内區塊，以對該視訊進行編碼。 如此來，如果要進行兩台螢幕同步顯干/ g 傳遞所須進行的影像壓縮技術，當可门‘:::動：： (jumon estlmatlon)的運算量 製螢幕操作晝面時，^ 右疋*要進订錄 算量也可改善錄Ϊ的;S技術改良其移動估測的運 【實施方式] 為讓本發日月之卜、+、 顯易僅，下文特舉出較佳實=的％特徵和優點能更明 詳細說明如下。 實施例，並配合所附圖式，作 的影像中’有一物體移 請參考圖1，其顯示在連續 1353792 動。如先刖技術所描述，利用人類的視覺暫留 作其實都是由一張張影像(frame)所構成的； 張連；的影像’第-影像η與第二影像12，而二兩〇 的移動，則分職示在第―料u與第三 =，也就是物體丨°是從第-影像11移動到 請同時參考圖3的流程圖，本發明所提 括以下步驟： 』々在匕 =1 詈該物體1〇分別在第一影像11與第二影像 之位置育訊。例如該些位置資訊可以 統（0S)。 作業糸 步驟S12:依據該物體1〇之位置資訊，計算含 料向量。請时參相1A的區 =屬=。較佳者’區塊屬性15包括晝面内區塊（i贈 〇C 及/或一畫面内區塊（intra block) 152。在本實many. When you restore, you can subtract the difference from the previous image to get the value of the next one. ‘ U and there must be moving objects in the continuous movie, that is, the positions of the two objects in the front and back are different. If you can find the object shift _ # position, then do subtraction' then the difference will be smaller. Therefore, it is necessary to find the motion vector, and the way to estimate where the object moves is the Motion Estimation. When restoring an image, finding the image block to subtract the difference from the motion vector is called Motion Compensation. The motion vector is simply estimating the current frame (for example, 8x8 or 16x16). Macro Block The vector of the reference block displacement that is most similar to the previous image (Previous Frame). Subtracting a certain block of the same size in the Macro Block and Reference Block that is to be used in the current image, and then adding the absolute value of each subtracted pixel to obtain a value called absolute difference sum. (SAD), where the smaller the SAD value, the closer the two blocks are. A SAD value is calculated for all positions in the search area, and the smallest value is the position after the movement to be searched, so that the motion vector can be calculated. Although the use of motion vectors and differences can reduce the amount of computational data to be stored, it is still a lot of computational effort to find the motion vector compared to two images. In addition, since the technical architecture of image compression is well known to those skilled in the art, it is not described here. 6 1.353792 SUMMARY OF THE INVENTION A primary object of the present invention is to provide a method of video encoding. Another object of the present invention is to provide a video editing that reduces the amount of computation of motion estimation in an image discarding technique. To achieve the above object, the present invention provides a method of video encoding. The video is composed of a plurality of frames, and includes a plurality of blocks. The video object has a position (four) method including the following steps: Step S11H takes the position information of the object-figure image and the second image. . For example, the location information can be from an operating system (0 s). Step S12: Calculate the area containing the object according to the position information of the object. Preferably, the U-block included in the second image is defined as an inter-block (inter_block), and is encoded according to the estimated motion vector. In this embodiment, the present invention can calculate the position or size of the object according to the position information of the object, and the corresponding position will be newly appeared in the first image due to the change of the position or size of the object. The block is defined as an intra-block for encoding. Step (1): encoding the video according to the estimated motion vector. = The position of the body at the first image includes an identification code, and the 5 疋 is used to identify the object. Therefore, by the knowledge provided by the operating system 7 1353792, the 'identifiable object' u knows which position information is to be retrieved, thus saving the amount of calculation. ‘The object has a window frame, and the position information of the operating system can be used to know the coordinate values of the object in the first image, that is, the coordinate values of the four corner positions of the window frame. As the well-known operating system can provide window information, there are many details about the coordinate value. There are many possible embodiments for moving an object from a first image to a second image, including the simple movement of the object and the scaling of the object. Therefore, if the object moves simply, the size of the object is substantially the same as the second image in the first image. If the object is zoomed, the size of the object ^the second image is different from the second image, and the size of the object is enlarged or reduced. In one embodiment, the video is an operation of the operating system and the location information of the object is provided by the operating system. In addition to the simple movement of the above objects and the scaling of objects, objects can be added or deleted. Therefore, the present invention also discloses a method for reducing the video coding of the estimated amount of motion estimation when adding an object. The video is composed of a plurality of frames, each of the images including a plurality of blocks, and the video is included in the video. An object has a new object information and a position information, and the method includes: Step S21 ··Capturing the newly added object information of the object. Preferably, the new object information is from an operating system (OS). _ Step S22: When the object does not exist in a first image, and is added to a second image of 8 1353792, the location information of the object in the second image is captured. Step S23: defining a block containing the object as an intra-picture block to encode the video. In addition, the present invention also discloses a method for reducing the video coding of the calculation amount of the motion estimation when the object is deleted, and the method includes the following steps: Step S31: deleting the object f. The preferred body information is from an operating system. Detachment Step S32: When the object exists in the -first image, and is deleted in the second male image, the position information of the object in the first image is captured. ^ Step S33: The alpha "represented as the intra-regional block" newly deleted after the corresponding object of the second shadow silly towel is deleted, so as to encode the video. To perform the image compression technology required for two screen synchro/g transfer, when the operation of the gate '::::: (jumon estlmatlon) is screened, ^ 疋* to order The amount of calculation can also improve the recording; S technology improves the estimation of its movement [Implementation] In order to make the date of the month, +, and the obvious, only the characteristics of the better = The advantages can be more clearly described in detail below. In the embodiment, and in conjunction with the drawings, there is an object in the image. Please refer to Figure 1, which is shown in the continuous 13537792. As described in the prior art, the human vision is temporarily used. The reservations are actually composed of a single frame; Zhang Lian; the image 'the first image η and the second image 12, and the movement of the two or two ,, the division is shown in the first material u Third=, that is, the object 丨° is moved from the first image 11 to the flowchart of FIG. 3, the present invention The following steps are as follows: 々 匕 匕 =1 詈 The object 1 育 is respectively incubated at the position of the first image 11 and the second image. For example, the location information can be unified (0S). Home 糸 Step S12: According to the object 1〇Location information, calculate the material vector. Please refer to phase 1A area = genus =. Preferred 'block attribute 15 includes inside the block (i give 〇 C and / or an intra-picture block ( Intra block) 152. In this case
^例中’依據晝面内區塊(inter bl〇ck)更計算出其移動向 罝（mcrnon vector)。因此在進行影像壓縮時移動估測 (m〇U〇n eStimati〇n)之運算可利用區塊屬性15的晝面内區 塊152所預先估算出的移動向量而無須--比對第Γ 影像11及第二影像12的所有區塊。 一請回到圖i與圖1A，物體1〇在第一影像u之位置 yK包括-識別碼’且該識別碼是用以辨識物體1〇 本實施例中，物體10具有一視窗框，利用作 業糸、先（圖未不）的位置資訊可獲知該物體1〇在該第—影 ^353792 像之四個座標值10a〜1〇d，也就是該視窗框之四個轉角 (corner)位置的座標值。 舉例而言，電腦的作業系統可提供物體1〇視窗框的 識別碼，因此藉由作業系統所提供的識別碼，即可辨識 物體10，以了解是要擷取哪一個物體10的位置資訊。例 如利用系統程式透過作業系統呼叫：In the example, the moving direction (mcrnon vector) is calculated based on the inter bl〇ck. Therefore, the operation of the motion estimation (m〇U〇n eStimati〇n) during image compression can utilize the motion vector estimated by the intra-block 152 of the block attribute 15 without the need to compare the second image. 11 and all blocks of the second image 12. Returning to FIG. 1 and FIG. 1A, the object 1 包括 includes the -identification code ' at the position yK of the first image u and the identification code is used to identify the object 1 . In the embodiment, the object 10 has a window frame, which utilizes The position information of the operation 先 and the first (not shown) can be obtained from the four coordinate values 10a~1〇d of the object 353792 image, that is, the four corner positions of the window frame. Coordinate value. For example, the operating system of the computer can provide the identification code of the window frame of the object, so that the object 10 can be identified by the identification code provided by the operating system to know which position information of the object 10 is to be retrieved. For example, using a system program to call through the operating system:
GetDesktopWindows()" 以取知物體1 0的識別碼，藉此更進一步可利用系 程式透過作業系統呼叫Ί hGetDesktopWindows()" to know the identification code of object 10, so that the program can be further called through the operating system.
GetWindowRect( )f, 以取得物體10視窗框的座標值。 如眾所周知者’作業系統可提供視窗資訊，因 於擷取位置資訊的技術未多贅述。 —實施例中，本發明之視訊為—作㈣統畫面之 刼作過程，且物體10之位置資訊是由作業系統 供。藉此，當使用者進行遠端桌面連線時Υ、傳送作業 統的操作畫面成為-視訊，此時通常會對此視訊進行編 ，屬縮後才傳送到遠端。故此時可利用作業系 的 資訊，減少此視訊編碼所需的運算。 八’、”、’、的 此外，物體從第一影像移動到第二影 考第圖2，在此實施例中，物體2。是從第： ^ 縮第一影像22的物體2〇，。同樣地，本發GetWindowRect( )f, to get the coordinate value of the object 10 window frame. As is well known, the operating system provides window information, as the technology for capturing location information is not repeated. - In the embodiment, the video of the present invention is a process of making a picture of the (four) system, and the location information of the object 10 is provided by the operating system. Therefore, when the user connects to the remote desktop, the operation screen of the transmission and operation system becomes a video, and the video is usually edited and transmitted to the remote end. Therefore, the information of the operating system can be utilized at this time to reduce the operations required for the video encoding. In addition, the object moves from the first image to the second image. In this embodiment, the object 2 is an object 2 from the first image 22. Similarly, this hair
Jr aa 、法可利用電腦的作業系統提供物體20、20，視窗 座標值。藉此，即可獲得如圖2八所示的區 5。在本實施例中，除了物體2〇盥2〇，之 與/二影像22僅差別在於物體2。縮小之後 定義為餘的空間在區塊屬性25中顯示即 法，：旦=，intrabl〇ck)，也就是依照本發明之方 塊。τ視m編碼時，多餘的空間會被定義為畫面内區 由以上圖1與圖2可知，若物體1〇僅 寸在該第-影像11與該第 在第目二，體20是進行縮放’則物體20、2〇’的尺寸 小物體^寸21。與第二影像22不同4分別為放大或縮 雜秘ί 了上述物體單純的移動及物體縮放之外，還可 :移動估測的運算量之視訊編輪法。請二:二 與圖5，新增㈣6 0在第二料6 2，也就圖 並無該物體60。在此實施例中，本發明之方法^像 二，S21 . #貞取物體6G之—新增物訊 增物體資訊是來自一作業系統。 專乂佳者新 Ϊ驟a/ Ϊ物體於一第一影像中不存在，而新增於-第-衫像時’擷取該物體在該第二影像中之該位置資訊。 步驟S23 :將該影像含有該物體之 塊’以對該視訊進行編碼。在本實施例中， 1353792 間區塊，也就是區塊屬性（圖未示）僅包含畫面内區塊（圖 未示）’因此視訊編碼是依據畫面内區塊。 同樣地，在本實施例中，物體6〇具有一視窗框，利 用作業系統的位置資訊可獲知該物體在該第一影像61之 四個座標值，也就是該視窗框之四個轉角（c〇rner)位置的 座標值。如眾所周知者’作業系統可提供視窗資訊，因 此關於擷取座標值的技術未多贅述。Jr aa, Fa can use the computer's operating system to provide objects 20, 20, window coordinates. Thereby, the zone 5 as shown in Fig. 2 can be obtained. In the present embodiment, except for the object 2〇盥2〇, the difference between the image and the image 2 is only the object 2. After the reduction, the space defined as the remainder is displayed in the block attribute 25, i.e., intrabl〇ck), that is, the block according to the present invention. When τ sees m coding, the extra space is defined as the inner region of the picture. As can be seen from Fig. 1 and Fig. 2 above, if the object 1 is only in the first image 11 and the second object, the body 20 is scaled. 'The size of the object 20, 2〇' is small object ^ inch 21. Different from the second image 22, respectively, the zooming or shortening of the simple movement of the object and the scaling of the object may be: a video stitching method of moving the estimated calculation amount. Please two: two with Figure 5, add (four) 60 in the second material 6 2, also the figure does not have the object 60. In this embodiment, the method of the present invention is as shown in Fig. 2, S21. #取取物6G-Additional material information The object information is from an operating system. Specialist New Step a/ The object does not exist in a first image, but is added to the -first-shirt image to capture the location information of the object in the second image. Step S23: the image is included in the block of the object to encode the video. In this embodiment, the block of 1353792, that is, the block attribute (not shown) contains only the intra-picture block (not shown), so the video coding is based on the intra-picture block. Similarly, in the embodiment, the object 6 has a window frame, and the position information of the operating system can be used to know the four coordinate values of the object in the first image 61, that is, the four corners of the window frame (c座rner) The coordinate value of the position. As is well known, the operating system provides window information, so the technique for extracting coordinate values is not described.
在一實施例中，該視訊為一作業系統畫面之操作過 程，且物體之位置資訊是由該作業系統提供。 ，^外，本發明亦揭露在刪除物體時，降低移動估測 的運算量之視訊編碼的方法。請同時參考圖6與圖7,從 η 81刪除物體80 ’也就是第二影像82並無該物 體80。在此實施例中，本發明之方法包含：In one embodiment, the video is an operating system screen operation and the location information of the object is provided by the operating system. In addition, the present invention also discloses a method of reducing the video coding of the calculation amount of the motion estimation when deleting an object. Referring to Figures 6 and 7, the object 80' is deleted from η 81, i.e., the second image 82 does not have the object 80. In this embodiment, the method of the present invention comprises:
==取來:趙二之系-:除物_。較佳者 擷取物體80在第一影 步驟S32 :依據該刪除物體資訊 像81之一位置資訊。 義Λ像中，對應於該物體刪除後所新出 内區塊’以對該視訊進行編碼。在 像82並無差別，因此其區塊屬 ;= 區塊（圖未示）〇 、口不丨”重巴3 I面内 由於上述之方法最好 以電腦程式完成，因此本發明 13 1353792 之方法可以以電腦程式寫於載有電腦可使用之媒介，聲 如硬碟，積體電路，光碟及其他可記錄電腦程式之物品: 腦而言’本發明之視訊編碼之方法，可以電 微便執行，而1腦程式可以儲存於任何 微處理早兀可以辨識、解讀之紀錄媒體，或 錄媒體之物品及裝置或電腦程式產品 =开 式，該物品較佳為CD，CD_R，M0，軟碟=為= =IC晶片或任何熟悉此項技藝者所可使 有 =媒=°。由於本發明之視訊編碼之方法= 二整牺任何熟悉電腦程式語言者閱讀本發明說 腦軟體程式’故有關電腦軟體程式細節部分 雖然本發明已以較佳實施例揭露如上，鈇其並 二：定J發明’任何熟習此技藝者，在不脫離本發明之 作些許之更動與潤掷，因此本= 保杨圍虽視後附之申請專利範圍所界定者為準。 【圖式簡單說明】 圖1係顯示—物體在連續的兩張影像中移動之示意圖。 圖1Α係依據圖1，顯示其區塊屬性。 圖2係顯示—物體在連續的兩張影像中縮放之示意圖。 圖2A係依據圖2，顯示其區塊屬性。 1353792 圖3係依據本發明之實施例，顯示降低移動估測（motion estimation)的運算量之視訊編碼的方法流程圖。 圖4係顯示一物體在連續的兩張影像中新增之示意圖。 圖5依據圖4,顯示其降低移動估測的運算量之視訊編碼 的方法流程圖。 圖6係顯示一物體在連續的兩張影像中被刪除之示意圖。 圖7依據圖6,顯示其降低移動估測的運算量之視訊編碼 的方法流程圖。 【主要零件符號說明】 物體 10、20、20’ 、40、60 ' 80 第一影像 11、21、41、61、81 第二影像 12、22、42、62、82 區塊屬性15、25 晝面間區塊（inter block)152、252 晝面内區塊（intra block)l51、251 15== Fetched: Zhao Erzhi -: Decontamination _. Preferably, the object 80 is captured in the first shadow step S32: according to the position information of the deleted object information image 81. In the image, the video is encoded corresponding to the new block after the object is deleted. There is no difference in the image 82, so the block belongs to it; = the block (not shown) is not in the mouth. In the face of the heavy bar, the above method is preferably completed by a computer program, so the present invention 13 1353792 The method can be written in a computer-readable medium, such as a hard disk, an integrated circuit, a compact disc and other items capable of recording a computer program: The brain of the present invention can be used for the method of video coding. Execution, and the 1 brain program can be stored in any recording medium that can be identified and interpreted by the micro-processing, or the recorded media and device or computer program product = open type, the item is preferably CD, CD_R, M0, floppy = = = IC chip or any one skilled in the art can make = media = °. Because of the video coding method of the present invention = two people who are familiar with the computer programming language to read the invention of the brain software program' Computer Software Program Details Although the present invention has been disclosed in the preferred embodiments as described above, it is a matter of course: "Anyone skilled in the art will be able to make some changes and tempering without departing from the invention, therefore Yang Wei is subject to the definition of the patent application scope attached. [Simplified illustration of the drawing] Figure 1 shows a schematic diagram of the movement of an object in two consecutive images. Figure 1 shows the block according to Figure 1. Figure 2 is a diagram showing the scaling of an object in two consecutive images. Figure 2A shows its block properties in accordance with Figure 2. Figure 1353 shows a reduced motion estimation (motion) in accordance with an embodiment of the present invention. Figure 4 is a flow chart showing the method of video encoding of an operation amount of Fig. 4. Fig. 4 is a schematic diagram showing an object added to two consecutive images. Fig. 5 shows the video coding of the operation amount for reducing the motion estimation according to Fig. 4. Method Flowchart Figure 6 is a schematic diagram showing the deletion of an object in two consecutive images. Figure 7 is a flow chart showing the method of reducing the video coding of the calculation of the motion estimation according to Figure 6. [Main part symbol description 】 object 10, 20, 20', 40, 60' 80 first image 11, 21, 41, 61, 81 second image 12, 22, 42, 62, 82 block attribute 15, 25 昼 inter-block ( Inter block) 152, 252 Intra block l51, 251 15