A Low Complexity Algorithm for H.264/AVC Intra Prediction

Intra prediction plays an important role in H.264/AVC. To reduce the complexity of the intra prediction of H.264/AVC, this paper proposes an efficient intra prediction algorithm. First, we determine if a macro block (MB) is flat or detailed rich by comparing its texture feature with dual thresholds criterion. Then, the flat MBs will be encoded in Intra 16 16 mode and those relatively rich MBs are encoded in Intra 4 4 mode. Second, an efficient Intra 4 4 mode selection algorithm is also proposed to reduce the number of candidates of intra prediction modes. The experimental results show that the encoding time of I-frame is reduced by an average of 30%, and both the increment in output bit rate and decrement in Peak Signal-to-Noise Ratio (PSNR) are negligible.