A novel post-processing method for H.264/AVC video quality enhancement

Lossy compression has been widely used to reduce the bandwidth for image/video transmission or storage, especially in networks in which high compression ratio is often expected. However, the coded video suffers from different degree of distortions. Post-processing can reduce these distortions and alleviate the conflict between bit rate reduction and quality preservation. In this paper we propose a linear minimum mean square error (LMMSE) based post-processing algorithm in DCT domain to enhance the video quality after compression. We treat the reconstructed DCT coefficients before deblocking and after deblocking as two noisy observations of the DCT coefficients of the original frame, and try to estimate the original ones by merging the two observations optimally. Experimental results demonstrate that our proposed algorithm can efficiently improve the video quality by about 0.5 dB.