在視訊壓縮標準H.264/AVC下的一個有效率畫面內編碼方法

Abstract

國際組織ISO/IEC和ITU-T共同制定了一套名為H.264/AVC的最新視訊壓縮標準。H.264/AVC可以達到比以往其他視訊壓縮標準更高的壓縮倍率，然而卻因此付出極多的壓縮時間。畫面內模式選擇在標準裡，對於4x4大小的區塊，提供了9種模式選擇，而對於16x16大小的區塊，提供了4種模式的選擇。在這篇論文裡，我們對於畫面內模式選擇提供了一套有效率的演算法。我們將會使用一種名為 ”快速過濾畫面內模式方法”，將一些模式成為候選模式，並且針對候選模式來做選擇。同時我們也使用一些空間上的資訊，使畫面內預測的演算法提早結束，達到加快編碼時間的效果。實驗結果顯示，我們的演算法在我們設定的編碼環境下，與暴力法搜尋比較能節省28.288%的編碼時間，且品質僅降低0.056dB，位元率僅上升0.939%。同時此結果也優越於Pan等人所發表的方法。

Two international organizations named ISO/IEC and ITU-T had developed the H.264/AVC video coding standard that is the newest one by now. Although H.264/AVC can achieve higher coding efficiency than the previous standards, its encoding time complexity is unbearable. In this thesis, we will present an efficient algorithm for the intra mode decision which has nine prediction modes for a 4x4 block coding, and four prediction modes for a 16x16 block coding. A Fast Intra-mode Filtering Method (FIFM) is provided to quickly find out the candidate modes, and the spatial coherence is utilized to achieve some earlier termination. Experimental results show that the proposed algorithm can reduce the time complexity about 28.288% with 0.056dB loss of PSNR and 0.939% increment of bit-rate comparing with the RDO full search scheme. This result also shows that the proposed method is superior to the algorithm proposed by Pan et. al. under the same encoding conditions.