Hierarchical Block Size Decision for H.264/AVC Intra Prediction

In H.264/AVC intra coding, rate-distortion optimization (RDO) is employed to determine the best block size. It demands a huge amount of computing resource, especially for the High Profile which incorporates the additional 8×8 block wise prediction. Obviously, block size prediction with high accuracy could reduce the computational complexity significantly. For this purpose, a two-stage block size decision scheme for H.264/AVC intra prediction is proposed in this work. At first stage, the sum of absolute adjacent pixel error (SAAPE) is used to classify the macroblock into texture and homogeneous groups. Then, the SAAPE of sub-blocks are further applied to determine the best block size in the second stage. With such two-stage design, only one or two block sizes for each macroblock is selected for subsequent RDO process. Extensive experimental results have shown that proposed algorithm leads to over 45% reduction in encoding time on average with negligible degradation in rate-distortion performance.

[1]  Homer H. Chen,et al.  Fast Decision of Block Size, Prediction Mode, and Intra Block for H.264 Intra Prediction , 2010, IEEE Transactions on Circuits and Systems for Video Technology.

[2]  Itu-T and Iso Iec Jtc Advanced video coding for generic audiovisual services , 2010 .

[3]  Liu Weihua,et al.  Fast intra mode decision algorithm in H.264/AVC using characteristics of transformed coefficients , 2008 .

[4]  Jhing-Fa Wang,et al.  Intensity Gradient Technique for Efficient Intra-Prediction in H.264/AVC , 2008, IEEE Transactions on Circuits and Systems for Video Technology.

[5]  G. Bjontegaard,et al.  Calculation of Average PSNR Differences between RD-curves , 2001 .

[6]  Kai-Kuang Ma,et al.  Hierarchical Intra Mode Decision for H.264/AVC , 2010, IEEE Transactions on Circuits and Systems for Video Technology.

[7]  Susanto Rahardja,et al.  Fast mode decision algorithm for intraprediction in H.264/AVC video coding , 2005, IEEE Transactions on Circuits and Systems for Video Technology.