PROBLEM TO BE SOLVED: To improve efficiency of quasi-static entropy encoding including Canonical Huffman encoding.;SOLUTION: A selector 1 divides source data M which is a sequence of a plurality of symbols according to the kind of the symbols. Each 1st CH encoder 2 outputs individual code words constituting a portion of a 1st code sequence M' one after another by carrying out Canonical Huffman encoding processing of the data divided by the selector 1 as input by using characteristic dictionary information corresponding to the kind of symbols associated with the 1st CH encoder itself. A 2nd CH encoder 3 outputs a 2nd code sequence E' and 2nd additional information E'' being dictionary information by performing Canonical Huffman encoding processing using, as input, 1st additional information E obtained by aggregating a plurality of pieces of dictionary information used by the plurality of 1st CH encoder 2.;COPYRIGHT: (C)2006,JPO&NCIPI
展开▼