A neural-net based rate control algorithm for MPEG encoders is introduced. It is based on the global rate-distortion model constructed by a self-organising feature map. The performances of the proposed algorithm are b...
详细信息
A neural-net based rate control algorithm for MPEG encoders is introduced. It is based on the global rate-distortion model constructed by a self-organising feature map. The performances of the proposed algorithm are better than those for MPEG-4 VM5+ frame-based ratecontrol, based on the criteria of the average number of bits per Frame and PSNR.
A rate control algorithm is presented for constrained variable bit rate (C_VBR) encoding to make the compressed video stream more friendly to the user network interface (UNI) than in free VER coding. Experiments show ...
详细信息
A rate control algorithm is presented for constrained variable bit rate (C_VBR) encoding to make the compressed video stream more friendly to the user network interface (UNI) than in free VER coding. Experiments show that the algorithm would not only meet the delay constraint and traffic contract of VER transfer so as to avoid the cell loss over UNI, bur also benefit the statistical multiplexing by reducing the burstiness of the MPEG encoded stream.
暂无评论