Grammar-based Compression using Multi-phase Hierarchical Segmentation

A. Akimov and P. Fränti (Finland)

Keywords

: Data compression, dictionary, grammar-based, hierarchical parsing, LZMW.

Abstract

We propose a multiphase hierarchical segmentation for data compression. The proposed method constructs a semi-adaptive dictionary by growing the dictionary iteratively using the LZMW algorithm. The result of the growing will be a complete partition of the input file into a binary tree structure. The tree is then pruned into a feasible size as it is also stored in the compressed file. The proposed method can be seen as a special case of grammar-based compression.

Important Links:



Go Back