近年來發展出愈來愈多能夠支援高解析度的多媒體產品，為了能夠有效壓縮伴隨著高解析度而來的龐大資料量，HEVC使用許多技術來進行壓縮。在HEVC標準架構中的畫面間預測(Inter Prediction)主要包含合併模式決策(Merge Mode Decision)與畫面間模式決策(Inter Mode Decision)。回顧先前研究，利用雙向絕對誤差和(Sum of Absolute Bi-Prediction Differences, SABPD)與模板匹配(Template Matching, TM)演算法，其利用已編碼完成的資訊來進行預測，在解碼端也能使用相同的資訊進行預測，如此便能減少需要傳送的標頭檔資訊，以節省傳送的位元率。為了提升編碼效能，本篇論文中提出擴充型合併模式決策(MergeExtended)，利用時間域上的候選區塊來尋找更準確的預測畫面。為了節省編碼時間，提出了畫面間模式提早決策(Early InterSABPD/TM)。最後將擴充型合併模式決策與畫面間模式提早決策做結合，由實驗結果得知其BDBR平均下降2.213%。;In recent years, lots of multimedia devices with high resolution are developed. In order to compress the significant increasing of data storage effectively, HEVC utilize multiple techniques to compress the data. Standard structure of HEVC includes both Merge Mode Decision and Inter Mode Decision in Inter Prediction. Review the previous work, using Sum of Absolute Bi-Prediction Differences(SABPD) and Template Matching(TM) algorithm, which utilizes encoded information to predict. Decoder can predict with the same information, so we can reduce transmit of header information to decrease bitrate. In this thesis, we proposed MergeExtended which using extended candidates on Temporal Domain in Merge Mode Decision to improve the encoding performance. We proposed Early InterSABPD/TM to reduce encoding time. Finally, we combine MergeExtended and Early InterSABPD/TM. According to experimental results, our proposed algorithm achieves the reduction of 2.213% BDBR
