System Level Design of the Adaptive Arithmetic Encoding Used in the JPEG 2000 Standard

Hardware implementation of the JPEG 2000 image compression standard for real-time image compression requires implementation of the adaptive arithmetic encoder among other related algorithms. In this work, a system level design for the adaptive arithmetic encoder used in the JPEG 2000 standard is proposed. This design is based on the assumption that the bit-sequence and its corresponding context, obtained through bit modelling for the encoder, are input to the system in their proper order. The output bit-sequence is also used as input to the encoder bit modelling. The original bit-sequence provides the context based on the assumption that the first pass is the cleanup pass and the first context is the Run Length context. All other contexts are recursively evaluated as arithmetic encoder encodes the bit sequence. The final encoding is lossless and results in perfect recovery of the original block image when proper decoding algorithm is used.