Code book construction for variable to variable length entropy encoding

摘要

A method of constructing a code book for groupings of symbols drawn from an alphabet, in which variable-sized groups of symbols are each assigned a variable length code based on probability of occurrence of symbol groupings. Code book entries are added by tentatively extending the high probability groupings with symbols from the alphabet. Code book size is restrained by identification of identify high probability symbol groupings, such that low probability groupings are combined into a single code book entry. Probability of occurrence for each entry is tracked. Extension and combination is repeated until a code book of predetermined size is reached. Each code book entry is assigned an entropy-type code according to the probability associated with each book entry.