HEVC視窗視訊編碼畫面內區塊複製技術之研究

Abstract

隨著網路與電腦的普及，電腦影像的壓縮與傳輸亦日漸重要。為了滿足其廣泛的應用需求，如遠端桌面，視訊會議，遠距教學，雲端遊戲與計算等，國際標準委員會Joint Collaborative Team on Video Coding正在制定視窗視訊編碼標準，作為高效率視訊編碼(High Efficiency Video Coding, HEVC)的延伸之一。 本論文研究HEVC視窗視訊編碼畫面內區塊複製(Intra Block Copy, IBC)之技術。其核心概念相似於畫面間移動補償，僅不同於使用當前畫面作為參考畫面。我們對IBC相關議題，如區塊向量的傳輸、編碼效能與時間複雜度、及所需的記憶體存取頻寬之間的取捨進行研究。我們特別關注在SCM-3.0標準軟體中的IBC雜湊搜索方法，藉由收集現行雜湊表的編碼數據資料，檢查其效率並探討進一步改進的可能性。我們提出篩選有效的雜湊節點與一改良的雜湊函式演算法。實驗結果顯示，將本算法整合至視窗視訊編碼測試平台SCM-3.0時，其減少了10％至70％的雜湊表的記憶體使用量，且幾乎對位元率和編解碼時間無影響。

With the popularity and development of digital terminals, the demands for screen content video compression and transmission grow up. To meet a wide range of needs for video applications such as remote desktop, video conference, distance education, cloud gaming and computing, the ISO/ITU Joint Collaborative Team on Video Coding committee is further specifying the Screen Content Coding (SCC) standard, as one of the extensions of High Efficiency Video Coding (HEVC). In this thesis, the adopted Intra Block Copy (IBC) coding tool for HEVC Screen Content Coding is studied. The concept of IBC is essentially the same as motion compensation, except that the current frame is used as the reference. Issues of IBC such as the block vector signaling, the trade-off between coding performance and time complexity, and the required memory fetching bandwidth are investigated. Particularly, we look into the IBC hash search method in SCM-3.0 test model. We collect the coded data using the current hash table and examine their efficiency and explore possible ways for further improvement. A scheme of selecting effective hash nodes and a modified hash key generation method are presented. Experimental results show that the proposed method reduces 10% to 70% hash table memory usage but preserve the similar BD-rate savings and encoding complexity, when integrated into the SCM-3.0 test model.