DMCA
Finite Automata Based Compression of Bi-level and Simple Color Images (1996)
| Citations: | 17 - 1 self |
BibTeX
@MISC{Culik96finiteautomata,
author = {Karel Culik and II and Vladimir Valenta},
title = {Finite Automata Based Compression of Bi-level and Simple Color Images},
year = {1996}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
We introduce generalized finite automata as a tool for specification of bi-level images. We describe an inference algorithm for generalized finite automata and a lossy compression system for bi-level and simple color images based on this algorithm and vector quantization. 1 Introduction A bi-level multiresolution image is specified by assigning the value 0 or 1 to every node of the infinite quadtree. If the outgoing edges of each node of the quadtree are labeled 0; 1; 2; 3, we get a uniquely labeled path to every node; its label is called the address of the node. The address of a node at depth k is a string of length k over the alphabet f0; 1; 2; 3g. Hence, a bi-level multiresolution image can be specified as a subset of strings over the alphabet f0; 1; 2; 3g, namely the collection of the addresses of the nodes assigned value 1 (black). Regular sets of strings are specified by finite automata or regular expressions [13]. Therefore, finite automata can be used to specify (regular) ...
