200952492 九、發明說明： 【發明所屬之技術領域】 本發明是有關於一種應用於移動估測之適應性搜尋框 的大小決定方法與其應用。 【先前技術】 H.264/AVC是國際電信聯盟遠程通信標準化組（ITU-T) 内的視訊編碼專家組（Video Coding Experts Group ; VCEG) 與國際標準組織（ISO)/國際電工協會（IEC)下的動態圖像專 家組（Moving Picture Experts Group ; MPEG)共同組成的聯 合視訊組（Joint Video Team ; JVT)所制訂的視訊壓縮標準。 其中H.264係依據ITU-T以H.26x系列為標準之命名，而 AVC則為ISO/IEC MPEG的稱呼。此標準的第一版最終草 案（Final Draft)已於2003年5月完成。 H.264/AVC的主要目標是希望在近似的視訊品質下， 能較先前的相關技術（比如MPEG-2或者H.263) ’提供更高 的壓縮率；同時，不需增加很多複雜的編解碼工具，使得 硬體難以實現。另一個目標是可適應性，即該視訊壓縮標 準可在很廣泛的範圍内使用（包含高位元率與低位元率的 視訊，以及不同的視訊解析度），並且能在各種網路和系統 上（如多媒體行動電話系統或高晝質電視）工作。 移動估測（motion estimation，ME)在H.264中是很重要 的部份。移動估測主要目的是消除影像在時間域上的冗 餘’使其在編解碼過程時只需記錄第一張畫面的資訊，而 其它畫面皆可以由移動過程中所紀錄的等資訊來重建，如 200952492 此一來可以不用儲存所有晝面的資訊，即可使得資訊量大 幅減少而達到影像壓縮的效果。一般的移動估測方法係以 每個晝面框中的一個區塊（macro block)做為最小估測單 位，每個區塊之大小可例如為16X16個像素。 目前已有相當多的移動估測方法被提出，例如T.Koga 等人所提出之三步搜尋（Three-Step Search，TSS)演算法，又 例如Lai-Man Po等人所提出之四步搜尋（Four-Step Search, FSS)演算法。然而，在習知技術中，使用固定大小的搜尋 框來搜尋較為匹配（matching)且相似度較高之區塊，此舉不 但容易浪費系統的運算資源，也可能導致影像的峰值信噪 比（Peak Signal to Noise Ratio ; PSNR)值下降0 因此，需要一種適應性搜尋框（adaptive searching window)與應用此適應性搜尋框之移動估測方法。 【發明内容】 因此，本發明之一方面係提供一種適應性搜尋框與決 定其大小之方法。 本發明之另一方面係提供一種應用適應性搜尋框之移 動估測方法。 根據本發明之一實施例，在此適應性搜尋框（adaptive searching window)之大小的決定方法中，首先，提供一目標 圖框，此目標圖框包含至少一已編碼區塊和一目標區塊。 接著，提供一參考圖框（reference frame)，此參考圖框包含 複數個參考區塊。然後，進行一閥值決定步驟，以計算至 少一已編碎區塊之一第一平均絕對差值總和（Average of 6 200952492200952492 IX. DESCRIPTION OF THE INVENTION: TECHNICAL FIELD OF THE INVENTION The present invention relates to a method for determining the size of an adaptive search box for mobile estimation and its application. [Prior Art] H.264/AVC is the Video Coding Experts Group (VCEG) and the International Standards Organization (ISO)/International Electrotechnical Commission (IEC) in the International Telecommunication Union Telecommunication Standardization Group (ITU-T). Under the Motion Picture Experts Group (MPEG), the Joint Video Team (JVT) has developed a video compression standard. Among them, H.264 is based on the ITU-T standard named H.26x series, and AVC is the name of ISO/IEC MPEG. The first draft of this standard, the Final Draft, was completed in May 2003. The main goal of H.264/AVC is to provide a higher compression ratio than the previous related technologies (such as MPEG-2 or H.263) under the approximate video quality; at the same time, there is no need to add a lot of complicated editing. The decoding tool makes the hardware difficult to implement. Another goal is adaptability, that is, the video compression standard can be used in a wide range (including high bit rate and low bit rate video, and different video resolution), and can be used on various networks and systems. Work (such as multimedia mobile phone systems or high-quality TVs). Motion estimation (ME) is an important part of H.264. The main purpose of the motion estimation is to eliminate the redundancy of the image in the time domain, so that only the information of the first picture needs to be recorded during the encoding and decoding process, and other pictures can be reconstructed by information recorded during the moving process. For example, 200952492 can save the image compression effect by saving all the information without saving all the information. The general motion estimation method uses a macro block in each frame as the minimum estimation unit, and each block can be, for example, 16×16 pixels. A number of mobile estimation methods have been proposed, such as the Three-Step Search (TSS) algorithm proposed by T. Koga et al., and the four-step search proposed by Lai-Man Po et al. (Four-Step Search, FSS) algorithm. However, in the prior art, a fixed-size search box is used to search for blocks that are more matching and have higher similarity, which is not only easy to waste the computing resources of the system, but also may cause the peak signal-to-noise ratio of the image ( Peak Signal to Noise Ratio; PSNR) value drops by 0 Therefore, an adaptive searching window and a mobile estimation method using this adaptive search box are needed. SUMMARY OF THE INVENTION Accordingly, one aspect of the present invention provides an adaptive search box and a method of determining its size. Another aspect of the present invention provides a motion estimation method using an adaptive search box. According to an embodiment of the present invention, in the method for determining the size of the adaptive searching window, first, a target frame is provided, the target frame includes at least one coded block and a target block. . Next, a reference frame is provided, which includes a plurality of reference blocks. Then, a threshold decision step is performed to calculate the sum of the first average absolute differences of at least one of the compiled blocks (Average of 6 200952492)
Sum of Absolute Difference)，並利用此第一平均絕對差值 總和來決定第一閥值和第二閥值，其中第二閥值係大於第 一閥值。接著’進行目標值計算步驟，以計算已編碼區塊 和目標區塊之一第二平均絕對差值總和。然後，進行一搜 尋框大小決定步驟，以比較第二平均絕對差值總和、第一 閥值和第二閥值之大小來獲得一比較結果，並根據此比較 結果來決定對應至目標區塊之搜尋框的大小。 根據本發明之另一實施例，此適應性搜尋框（adaptive ❹ searching window)之大小的決定方法中，首先，提供一目標 圖框’此目標圖框包含一目標區塊。接著，提供一參考圖 框（reference frame) ’此參考圖框包含複數個參考區塊。然 後’進行一閥值決定步驟’以計算這些參考區塊之一第一 平均絕對差值總和（Average of Sum of Absolute Difference)’並利用此第一平均絕對差值總和來決定第一閥 值和第二閥值’其中第二閥值係大於第一閥值。接著，進 行目標值計算步驟，以計算這些參考區塊和目標區塊之一 〇 第二平均絕對差值總和。然後，進行一搜尋框大小決定步 驟’以比較第二平均絕對差值總和、第一閥值和第二閥值 之大小來獲得一比較結果，並根據此比較結果來決定對應 至目標區塊之搜尋框的大小。 根據本發明之又一實施例，在此應用適應性搜尋框之 移動估測方法中，首先進行一閥值決定步驟，以計已編碼 區塊之一第一平均絕對差值總和，並利用第一平均絕對差 值總和來決定一第一閥值和一第二閥值，其中第二閥值係 大於第一閥值。然後，進行目標值計算步驟，以計算已編 7 200952492 碼區塊和目標區塊之—第二平均絕對差值總和。接著，進 行搜尋框大小決定㈣，職較第二平均絕對差值總和、 第一閥值和第二閥值之大小來獲得-比較結果，並根據比 較結果來決定對應至目標區塊之第—搜尋框之大小。然 後’進行相似度比較步驟，以利用第-搜尋框來於參考圖 框中選定匹配區塊。接著，進行移動向量計算步驟，以根 據目標區塊和匹8£區塊來計算目標區塊之移動向量。 〇 【實施方式】 在移動估測技術巾，若要使用適應性搜尋框（adaptive Searchingwindow)來搜尋較為匹配（matching)之區塊，適應 性搜尋框（adaptive seaixhing windGW)A小的蚊方法是不 可或缺的在以下的說明中，將以多個實施例來說明本發 明之適應性搜尋框大小的決定方法，與應用此適應性搜尋 框之移動估測方法。 請同時參照第1圖和第2圖，第1圖係搶示根據本發 ❹ 明第一實施例之目標圖框100和參考圖框150之結構示意 圓，其中相同之剖面線係代表相同之區塊（102、104和 106)。第2圖係繪示根據本發明第一實施例之適應性搜尋 框（adaptive searching window)大小的決定方法2〇〇的流程 不意圖。目標圖框100包含已編碼區塊1〇2、未編碼區塊 1〇4和處理令之目標區塊106。參考圖框15〇和目標圓框ι〇〇 為連續之圖框，其中參考圖框150包含有參考區塊152。 為決定目標區塊106於參考圖框150之搜尋範圍，進 行適應性搜尋框大小的決定方法200來決定目標區塊ι〇6 8 200952492 所對應之適應性搜尋框的大小。在決定方法2〇〇中，首先 進行閥值決定步驟210〇在閥值決定步驟21〇中，計算已編 碼區塊之一第一平均絕對差值總和（Average 〇f Sum 〇f Absolute Difference)，並利用第一平均絕對差值總和來決定 第一閥值和第二閥值。在本實施例中，第一閥值等於第一 平均絕對差值總和，而第二閥值等於第一閥值之12倍，在 本發明之其他實施例中，第一平均絕對差值總和與第一閥 值和第二閥值之關係可設定為其他比例。接著，進行目標 〇 值計算步驟220。在目標值計算步驟220中，計算已編碼區 塊和目標區塊之一第二平均絕對差值總和。然後，進行搜 尋框大小決定步驟230。在搜尋框大小決定步驟23〇中，比 較第一平均絕對差值總和、第一閥值和第二閥值之大小來 獲得比較結果，並根據比較結果來決定對應至目標區塊1〇6 之搜尋框的大小。在搜尋框大小決定步驟230中，當第二 平均絕對差值總和小於或等於第一閥值時，決定搜尋框之 大小為8乘8個參考區塊。當第二平均絕對差值總和大於 φ 第一閥值’且小於第二閥值時，決定搜尋框之大小為16乘 16個參考區塊。當第二平均絕對差值總和大於或等於第二 閥值時，決定搜尋框之大小為32乘32個參考區塊。 由以上說明可知，本發明之第一實施例係利用第二平 均絕對差值總和、第一閥值和第二閥值來預測目標區塊1〇6 之匹配區塊（matching block)可能出現的範圍，並且可動態 地調整搜尋框之大小’以在較小的計算複雜度下，達到較 佳的影像品質。 請參照第3圖，其係繪示根據本發明第二實施例之移 9 200952492 測方法300的流程示意圖。移動估測方法細係應用 適應性搜尋框大小的決定方法鳩’來計算目標區塊1〇6 ΟSum of Absolute Difference), and using the sum of the first average absolute differences to determine a first threshold and a second threshold, wherein the second threshold is greater than the first threshold. Next, a target value calculation step is performed to calculate a sum of the second average absolute differences of one of the coded block and the target block. Then, a search box size determining step is performed to compare the sum of the second average absolute difference, the first threshold and the second threshold to obtain a comparison result, and according to the comparison result, the corresponding to the target block is determined. The size of the search box. According to another embodiment of the present invention, in the method of determining the size of the adaptive ❹ searching window, first, a target frame is provided. The target frame includes a target block. Next, a reference frame is provided. This reference frame contains a plurality of reference blocks. Then 'perform a threshold decision step' to calculate the "Average of Sum of Absolute Difference" and use the sum of the first average absolute differences to determine the first threshold and The second threshold 'where the second threshold is greater than the first threshold. Next, a target value calculation step is performed to calculate the sum of the second average absolute difference of one of the reference block and the target block. Then, a search box size determining step is performed to compare the second average absolute difference sum, the first threshold and the second threshold to obtain a comparison result, and according to the comparison result, the corresponding to the target block is determined. The size of the search box. According to still another embodiment of the present invention, in the mobile estimation method for applying the adaptive search box, a threshold value determining step is first performed to calculate the sum of the first average absolute differences of one of the coded blocks, and utilize the A sum of average absolute differences determines a first threshold and a second threshold, wherein the second threshold is greater than the first threshold. Then, a target value calculation step is performed to calculate the sum of the second average absolute difference of the coded block 2009523492 and the target block. Then, the search box size decision (4) is performed, and the second average absolute difference sum, the first threshold value and the second threshold value are used to obtain a comparison result, and the corresponding to the target block is determined according to the comparison result. The size of the search box. Then, the similarity comparison step is performed to select the matching block in the reference frame by using the first search box. Next, a motion vector calculation step is performed to calculate a motion vector of the target block based on the target block and the block. 〇[Embodiment] In the mobile estimation technology towel, if an adaptive searching window is used to search for a matching block, the adaptive seaixhing windGW A small mosquito method is not available. In the following description, the method for determining the size of the adaptive search box of the present invention and the method for estimating the motion of the adaptive search box will be described in various embodiments. Referring to FIG. 1 and FIG. 2 simultaneously, FIG. 1 is a schematic diagram showing the structure of the target frame 100 and the reference frame 150 according to the first embodiment of the present invention, wherein the same hatching represents the same. Blocks (102, 104, and 106). Fig. 2 is a flow chart showing the method of determining the size of an adaptive searching window according to the first embodiment of the present invention. The target block 100 includes the encoded block 1〇2, the uncoded block 1〇4, and the target block 106 of the processing order. Reference frame 15 〇 and target circle frame ι 〇〇 are continuous frames, wherein reference frame 150 includes reference block 152. To determine the search range of the target block 106 in the reference frame 150, an adaptive search box size determination method 200 is performed to determine the size of the adaptive search box corresponding to the target block ι〇6 8 200952492. In the determining method 2, first, the threshold determining step 210 is performed, and in the threshold determining step 21, the first average absolute difference sum of the coded blocks is calculated (Average 〇f Sum 〇f Absolute Difference), And using the sum of the first average absolute differences to determine the first threshold and the second threshold. In this embodiment, the first threshold is equal to the sum of the first average absolute differences, and the second threshold is equal to 12 times the first threshold. In other embodiments of the invention, the sum of the first average absolute differences is The relationship between the first threshold and the second threshold can be set to other ratios. Next, target 〇 value calculation step 220 is performed. In target value calculation step 220, a sum of the second average absolute differences of one of the encoded block and the target block is calculated. Then, a search box size decision step 230 is performed. In the search box size determining step 23, comparing the sum of the first average absolute difference, the first threshold and the second threshold to obtain a comparison result, and determining the corresponding to the target block 1〇6 according to the comparison result. The size of the search box. In the search box size decision step 230, when the sum of the second average absolute differences is less than or equal to the first threshold, the size of the search box is determined to be 8 by 8 reference blocks. When the sum of the second average absolute differences is greater than φ the first threshold & is less than the second threshold, the size of the search box is determined to be 16 by 16 reference blocks. When the sum of the second average absolute differences is greater than or equal to the second threshold, the size of the search box is determined to be 32 by 32 reference blocks. As can be seen from the above description, the first embodiment of the present invention utilizes the second average absolute difference sum, the first threshold, and the second threshold to predict a possible matching block of the target block 1〇6. Range, and the size of the search box can be dynamically adjusted to achieve better image quality with less computational complexity. Please refer to FIG. 3, which is a flow chart showing the method 300 of the measurement method according to the second embodiment of the present invention. The method of determining the size of the adaptive search box 鸠' to calculate the target block 1〇6 Ο
，向量。在移動估測方法3〇〇中’首先進行適應性搜 哥框大小的決定方法2〇〇，以決定目標區塊ι〇6所對應之一 第一搜尋框之大小。接著，進行相似度比較步驟31〇。在相 似度比較步驟310 t，利用第-搜尋框來於參考圖框15〇 令選定一匹配區塊，此匹配區塊為第一搜尋框中，與目標 區塊106相似度最高之區塊’進行移動向量計算步 驟320。在移動向量計算步驟32〇中，根據目標區塊和匹配 區塊來計算目標區塊之移動向量。 經由複雜的實驗證明，在相同影像品質的條件下，本 發明之實施例的移動估測方法可較習知技術降低約6〇%的 計算複雜度，因此本發明之實施例的移動估測方法可節省 相當多的計算資源。 請參照第4圖，其係繪示根據本發明第三實施例之相 似度比較步驟310的流程示意圖。在相似度比較步驟31〇 中，首先，進行搜尋框比較步驟310a。在搜尋框比較步驟 31〇a中，判斷目標區塊106所對應之第一搜尋框是否大於 已編碼區塊102之最後一者所對應之一第二搜尋框，並提 供一判斷結果。上述之已編碼區塊102之最後一者係指目 標區塊106之前一區塊。例如若目標圖框〗〇〇之處理順序 為由左而右逐列處理，則目標區塊1〇6之前一區塊為目標 區塊106左方之區塊，又例如：若目標圖框1〇〇之處理順 序為由上而下逐行處理，則目標區塊106之前一區塊為目 標區塊106上方之區塊。 200952492 當判斷結果為是時’則進行搜尋步驟3i〇b，以利用第 -搜尋框和第二搜尋框之間的區域，來於參考圖框⑽中 搜尋目標區塊106之匹配區塊。#判斷結果為否時則進 行搜尋步称3U)。，以第—搜尋框來於參考圖框15〇甲 搜尋目標區塊106之匹配區塊。 在本發明第三實施例中，當判斷結果為是時，表示目 標區塊106與目標區塊106之前一區塊有相當的差異性， 目標區塊106之匹配區塊應會落在第—搜尋框和第二搜尋,vector. In the mobile estimation method 3, the method of determining the size of the adaptive search box is first performed to determine the size of the first search box corresponding to the target block ι6. Next, the similarity comparison step 31〇 is performed. In the similarity comparison step 310 t, a matching block is selected by referring to the frame 15 by using the first search box, and the matching block is the block with the highest similarity to the target block 106 in the first search box. A motion vector calculation step 320 is performed. In the motion vector calculation step 32, the motion vector of the target block is calculated based on the target block and the matching block. It has been proved by complicated experiments that the motion estimation method of the embodiment of the present invention can reduce the computational complexity of about 6〇% compared with the prior art under the condition of the same image quality, and therefore the motion estimation method of the embodiment of the present invention. Can save a lot of computing resources. Please refer to FIG. 4, which is a flow chart showing the similarity comparison step 310 according to the third embodiment of the present invention. In the similarity comparison step 31, first, a search box comparison step 310a is performed. In the search box comparison step 31a, it is determined whether the first search box corresponding to the target block 106 is greater than one of the second search boxes corresponding to the last one of the encoded blocks 102, and provides a determination result. The last of the above-described encoded blocks 102 refers to a block before the target block 106. For example, if the processing order of the target frame is processed from left to right column by column, the block before the target block 1〇6 is the block to the left of the target block 106, for example: if the target frame 1 The processing sequence is processed from top to bottom, and the previous block of the target block 106 is the block above the target block 106. 200952492 When the judgment result is YES, a search step 3i〇b is performed to search for the matching block of the target block 106 in the reference frame (10) by using the area between the first search box and the second search box. # When the judgment result is no, the search step is called 3U). The search block of the target block 106 is searched for by the first search box in the reference frame 15 armor. In the third embodiment of the present invention, when the determination result is YES, it indicates that the target block 106 is quite different from the previous block of the target block 106, and the matching block of the target block 106 should fall on the first- Search box and second search
框之間的區域，因此直接於第一搜尋框和第二搜尋框之間 的區域中搜尋目標區塊106之匹配區塊，以節省計算複^ 度0 ' 請同時參照第5圖和第6圖，第5圖係綠示根據本發 明第四實施例之目標圖框500和參考圖框55〇的結構示意 圖，第6圖係繪示根據本發明第四實施例之適應性搜尋框 大小的決定方法600的流程示意圖。目標圖框5〇〇包含未 編碼區塊504和處理中之目標區塊5〇6。在目標圖框5〇〇 〇 中，目標區塊506為第一個接受移動估測處理之區塊。參 考圓框550和目標圖框5〇〇為連續之圖框，其中參考圖框 550包含有參考區塊552。在決定方法6〇〇中首先，進行 闕值決定步驟610。在閥值決定步驟61〇中，計算參考區塊 552之一第一平均絕對差值總和，並利用此第一平均絕對差 值總和來決定一第一閥值和一第二閥值。在本實施例中， 第一閥值等於第一平均絕對差值總和，而第二閥值等於第 一閥值之1.2倍。接著，進行目標值計算步驟62〇。在目標 值計算步驟620中，計算參考區塊552和目標區塊5〇6之 11 200952492 -第二平均絕對差值總和。然後，進行搜尋框大小決定步 驟630。在搜尋框大小決定步驟630中，比較第二平均絕對 差值總和帛—閥值和該第二閥值之大小來獲得比 果’並根據比較結果來決定對應至目標區塊5〇6之搜尋框 的大小。在搜尋框大小決定步驟630中，當第二平均絕對 差值總和小於或等於第一閥值時，決定搜尋框之大小為8 乘8個參考區塊。當第二平均絕對差值總和大於第一閱值， 且小於第二閥值時’決定搜尋框之大小為16乘16個參考 〇 區塊。當第二平均絕對差值總和大於或等於第二閥值時， 決定搜尋框之大小為32乘32個參考區塊。 在本發明第四實施例中，目標區塊506為目標圖框5〇〇 中，第一個被處理之區塊，因此對於目標區塊5〇6而言， 目標圖框500中不存在有已編碼區塊。因此，本發明第四 實施例係利用參考圖框55〇的所有參考區塊，來決定第一 閥值和第二閥值。 雖然本發明已以實施例揭露如上，然其並非用以限定 ❹ 本發明’任何熟習此技藝者’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】 為讓本發明之上述和其他目的、特徵、和優點能更明 顯易懂’上文特舉一較佳實施例，並配合所附圖式，作詳 細說明如下： 第1圖係繪示根據本發明第一實施例之目標圖框和參 12 200952492 考圖框之結構示意圖。 第2圖係繪示根據本發明第一實施例之適應性搜尋框 (adaptive searching window)大小的決定方法的流程示意 圖。 第3圖係繪示根據本發明第二實施例之移動估測方法 的流程示意圖。 第4圖係繪示根據本發明第三實施例之相似度比較步 驟的流程示意圖， φ 第5圖係繪示根據本發明第四實施例之目標圖框和參 考圖框的結構示意圖。 第6圖係繪示根據本發明第四實施例之適應性搜尋框 大小的決定方法的流程示意圖。 【主要元件符號說明】 102 :已編碼區塊 106 :目標區塊 152 :參考區塊 210 :閥值決定步驟 230 :搜尋框大小決定步驟 310 :相似度比較步驟 310a :搜尋框比較步驟 310c :搜尋步驟 504 :未編碼區塊 550 :參考圖框 600 :大小決定方法 100 :目標圖框 104 :未編碼區塊 ❿ 150 :參考圖框 200 :大小決定方法 220 :目標值計算步驟 300 :移動估測方法 320 :移動向量計算步驟 31 Ob :搜尋步驟 500 :目標圖框 506 :目標區塊 552 :參考區塊 13 200952492 610 630 ： 閥值決定步驟 620:目標值計算步驟 搜尋框大小決定步驟The area between the frames, so that the matching block of the target block 106 is searched directly in the area between the first search box and the second search box to save the calculation complex 0 ' Please refer to FIG. 5 and FIG. 6 at the same time. Figure 5 is a schematic view showing the structure of the target frame 500 and the reference frame 55A according to the fourth embodiment of the present invention, and Figure 6 is a view showing the size of the adaptive search frame according to the fourth embodiment of the present invention. A schematic diagram of the process of determining method 600. The target frame 5〇〇 contains the uncoded block 504 and the target block 5〇6 in process. In target frame 5〇〇, target block 506 is the first block to accept the motion estimation process. The reference circle 550 and the target frame 5A are continuous frames, wherein the reference frame 550 includes a reference block 552. In decision method 6, first, a threshold decision step 610 is performed. In the threshold decision step 61, a sum of the first average absolute differences of one of the reference blocks 552 is calculated, and a first threshold and a second threshold are determined using the sum of the first average absolute differences. In this embodiment, the first threshold is equal to the sum of the first average absolute differences and the second threshold is equal to 1.2 times the first threshold. Next, the target value calculation step 62 is performed. In the target value calculation step 620, the sum of the reference block 552 and the target block 5〇6 200952492 - the second average absolute difference is calculated. Then, a search box size decision step 630 is performed. In the search box size determining step 630, comparing the sum of the second average absolute difference sum 阀-threshold and the second threshold to obtain a result 'and determining the search corresponding to the target block 5〇6 according to the comparison result The size of the box. In the search box size decision step 630, when the sum of the second average absolute differences is less than or equal to the first threshold, the size of the search box is determined to be 8 by 8 reference blocks. When the sum of the second average absolute differences is greater than the first read value and less than the second threshold, the size of the search box is determined to be 16 by 16 reference 〇 blocks. When the sum of the second average absolute differences is greater than or equal to the second threshold, the size of the search box is determined to be 32 by 32 reference blocks. In the fourth embodiment of the present invention, the target block 506 is the first processed block in the target frame 5, so that for the target block 5〇6, there is no target frame 500. Encoded block. Therefore, the fourth embodiment of the present invention uses the reference blocks of reference frame 55 to determine the first threshold and the second threshold. Although the present invention has been disclosed in the above embodiments, it is not intended to limit the invention, and the invention may be modified and modified without departing from the spirit and scope of the invention. The scope of protection is subject to the definition of the scope of the patent application attached. BRIEF DESCRIPTION OF THE DRAWINGS The above and other objects, features, and advantages of the present invention will become more <RTIgt; </ RTI> <RTIgt; </ RTI> <RTIgt; </ RTI> <RTIgt; The figure shows the structure of the target frame according to the first embodiment of the present invention and the structure of the reference frame of the reference 12 200952492. Fig. 2 is a flow chart showing a method of determining the size of an adaptive searching window according to the first embodiment of the present invention. Figure 3 is a flow chart showing a mobile estimation method according to a second embodiment of the present invention. 4 is a flow chart showing a similarity comparison step according to a third embodiment of the present invention, and FIG. 5 is a schematic structural view showing a target frame and a reference frame according to a fourth embodiment of the present invention. Figure 6 is a flow chart showing a method for determining the size of an adaptive search box according to a fourth embodiment of the present invention. [Main Component Symbol Description] 102: Encoded Block 106: Target Block 152: Reference Block 210: Threshold Determination Step 230: Search Box Size Determination Step 310: Similarity Comparison Step 310a: Search Box Comparison Step 310c: Search Step 504: Uncoded block 550: Reference frame 600: Size decision method 100: Target block 104: Uncoded block ❿ 150: Reference frame 200: Size decision method 220: Target value calculation Step 300: Motion estimation Method 320: Motion Vector Calculation Step 31 Ob: Search Step 500: Target Block 506: Target Block 552: Reference Block 13 200952492 610 630: Threshold Decision Step 620: Target Value Calculation Step Search Box Size Determination Step