A fast bitplane combination algorithm for bitplane coded scalable image/video

Bitplane coding techniques have been proposed as an effective approach to providing a bitstream truncation ability. They have also been adopted into the latest image and video coding standards, such as JPEG 2000, MPEG-4 and the emerging scalable video codec in MPEG-21, where bitplane coding techniques are used to generate a scaleable bitstream for various bitrates. However, bitplane decoding requires lots of bitwise operations while combining bitplanes into coefficients in the decoder. As a result, the straightforward combining algorithm is very inefficient and could become a bottleneck while decoding a high quality enhancement bitstream with several bitplanes. We propose an efficient bitplane combining algorithm which not only greatly speeds up the combining time by a factor of more than eight, but also makes the decoding time independent of the number of bitplanes to be combined. In addition, it is very suitable for VLSI implementation due to the parallel and table-lookup based structure.