Hardware-oriented Optimized Algorithm For EBCOT T1

JPEG2000 is a newly proposed next-generation still-image-compression standard. It not only has better compression performance than the existing JPEG standard, but also provides many new features, such as quality scalability, region of interest (ROI) coding. However, the computational complexity of JPEG2000 is much higher than that of JPEG. In order to satisfy the hardware implementation of a high efficient JPEG2000 system, we should simplify the original JPEG2000 algorithm.Discrete wavelet transform(DWT) and embedded block coding with optimized truncation (EBCOT) are the two main techniques adopted in JPEG2000. EBCOT is a two-tiered coder, where T1 is for bit-plane coder and a context-base adaptive arithmetic coder (MQ coder), and T2 is for rate-distortion optimization and bit stream layer formation.After detailed analysis of EBCOT T1, the coding information of each bit plane and the corresponding passes can be obtained simultaneously. Therefore, bit plane-parallel and pass-parallel coding algorithm is proposed to enhance the parallelism and increase the speed.An optimized algorithm for MQ encoder suitable for hardware implementation is presented in this paper. By reducing the complexity of the algorithm, the difficulty of hardware implementation for MQ encoder is reduced. Meanwhile, MQ encoder applied to pass-parallel is analyzed. MQ encoder is terminated on each coding pass, which strengthens the ability of error resilience. A three-stage pipelined architecture is used to enhance the parallelizability of the JPEG2000 system, the average processing speed can reach to 1bit/cycle.Finally, the test results show that the improved EBCOT T1 algorithm is suitable for hardware implementation. This method could improve the coding efficiency greatly while keeping the excellent performance of JPEG2000.