AN EFFICIENT ENCODING OF DCT BLOCKS WITH BLOCK-ADAPTIVE SCANNING

In a hybrid coder which employs motion compensation and discrete cosine transform (MC-DCT coder), up to 90 % of bits are used to represent the quantized DCT blocks. So it is most important to represent them with as few bits as possible. In this paper, we propose an efficient method for encoding the quantized DCT blocks of motion compensated prediction (MCP) errors, which adaptively selects one of a few scanning patterns. The scanning pattern selection of an MCP error block is based on the motion compensated images which are always available at the decoder as well as at the encoder. No overhead information for the scanning patterns needs to be transmitted. Simulation results show that the average bit race reduction amounts to 5 %.