Improved throughput arithmetic coder for JPEG2000

Increasing the throughput of the JPEG2000 block coder requires bit-plane and arithmetic coders capable of concurrent symbol processing. Previously described pipelined MQ coders are capable of consuming 1 symbol or less per clock cycle. We develop a new pipelined MQ coder that can process exactly two symbols per clock cycle. The technique is implemented on a FPGA, and is compared with our "Hypothesis Testing" arithmetic coder and a reference one symbol per cycle coder. Our implementation gives an increase in throughput of 1.9 times, at the cost of 1.7 times as much hardware, when compared to the reference coder. It also has 1.2 times the throughput, while consuming only 70% of the hardware associated with the Hypothesis Testing coder.

[1]  Liang-Gee Chen,et al.  Analysis and architecture design of block-coding engine for EBCOT in JPEG 2000 , 2003, IEEE Trans. Circuits Syst. Video Technol..

[2]  Gadiel Seroussi,et al.  Embedded block coding in JPEG2000 , 2000, Proceedings 2000 International Conference on Image Processing (Cat. No.00CH37101).

[3]  P. Glenn Gulak,et al.  Architectural advances in the VLSI implementation of arithmetic coding for binary image compression , 1994, Proceedings of IEEE Data Compression Conference (DCC'94).

[4]  Chen-Yi Lee,et al.  A high throughput low cost context-based adaptive arithmetic codec for multiple standards , 2002, Proceedings. International Conference on Image Processing.

[5]  Chein-Wei Jen,et al.  High-speed memory-saving architecture for the embedded block coding in JPEG2000 , 2002, 2002 IEEE International Symposium on Circuits and Systems. Proceedings (Cat. No.02CH37353).

[6]  Glen G. Langdon,et al.  An Overview of the Basic Principles of the Q-Coder Adaptive Binary Arithmetic Coder , 1988, IBM J. Res. Dev..