Computation bounded Video Coding

A realistic approach to coding algorithms has to assume that both computation and transmission powers are limited, while image quality has to be maximized. This is of paramount importance in the case of digital HDTV, where the source bit rate is extremely high. Packetized video transmission has the desirable feature of compensating for the source activity bursts, allowing a reduction in the computation requirements posed by the coding algorithms. Aim of the research is to evaluate the trade-offs between the transmission bit rate and the computation power required by a well established replenishment block coding algorithm for standard TV. The results of this research can easily be extended to HDTV. Conditional replenishment is controlled by a threshold on the block luminance difference energy, while computation bounds set the number of blocks processed in each frame. The actual threshold is determined by the input video buffer which accommodates the queuing blocks to be processed. Video quality is preserved using PCM, when the input buffer tends to overflow. The proposed coding strategy could be extended to computation and transmission bounded video compression. In this case both the input video buffer and the line buffer control the replenishment threshold, while the line buffer also controls the code quantizer.