隨著科技的進步，為了因應高畫質(HD)及超高畫質(Ultra HD)的影像需求，一個新的視訊編碼標準稱為高效率視訊編碼(High Efficiency Video Coding,HEVC)被訂定出來，以達到更好的編碼效能。HEVC的編碼單元(Coding Unit , CU)採用了四分樹(Quad-Tree)的編碼架構增加其編碼效能，在畫面內預測中，為了使預測更精準，使用了35個模式來進行預測，同時也大幅的增加其編碼的複雜度。因此本篇論文提出一低複雜度畫面內預測演算法，利用低頻交流值(Low-frequency AC value)及變異數(Variance)影像特徵值於編碼單元決策其深度的切割；於畫面內預測中，藉由變異數及梯度邊界特徵來判斷畫面同質性及預測單元(Prediction Unit,PU)的方向，減少約略模式決策(Rough Mode Decision, RMD)及RDO(Rate-Distortion Optimization)的候選模式個數。由實驗結果可知，我們提出的演算法在維持一定影像品質之下，平均能夠節省42%的整體編碼時間。;Along with the progress of science and technology, a new video coding standard called HEVC(High Efficiency Video Coding) is developed in order to achieve better coding performance for the emerging HD and Ultra HD video content. The coding unit(CU) of HEVC uses a Quadtree-based coding structure to increase coding performance. In order to predict more accurately, using 35 prediction modes in intra prediction. Simultaneously, it also increases its coding complexity. Hence, in this thesis, we proposed a low-complexity Intra prediction algorithm.We determinate the split of the depth of CU by using low-frequency AC value and Variance. In intra prediction , we use Variance and gradient boundary feature to distinguish the homogeneousness of image and direction of PU(Prediction Unit). Then, reducing the candidates of rough mode decision and rate-distortion optimization. The experimental results show that our proposed algorithm achieves 42% encoding time saving on average without significant degradation of coding performance.
