High-speed statistical compression using self-organized rules and predetermined code tables

Statistical compression methods using arithmetic coding have a high-compression ratio. However, their complexity limits their application due to the inherent computation overhead for a software implementation. There is an increasing need for faster data compression methods to improve the capabilities of existing data storage and communications technology. We have examined a statistical method which focuses on speed rather than compression performance. To obtain both a good compression ratio and fast processing speed, we previously proposed a simple statistical compression scheme using high-order context modeling and splay coding. However, splay coding still had a large processing overhead, even though it used binary code tree conversion instead of complicated statistical calculations. We aim to enhance the processing speed of our scheme while maintaining its high compression performance