Premium
A multi‐threshold adaptive filtering—an al approach to image enhancement
Author(s) -
Qian Jianzhong,
Yu KaiBor
Publication year - 1990
Publication title -
international journal of intelligent systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.291
H-Index - 87
eISSN - 1098-111X
pISSN - 0884-8173
DOI - 10.1002/int.4550050305
Subject(s) - filter (signal processing) , adaptive filter , computer science , noise (video) , image (mathematics) , context (archaeology) , artificial intelligence , computer vision , function (biology) , bilateral filter , kernel adaptive filter , filter design , pattern recognition (psychology) , mathematics , algorithm , paleontology , evolutionary biology , biology
There is a compromise between noise removal and texture preservation in image enhancement. It is difficult to perform image enhancement, using only one simple filter, for a real world image which may consist of many different regions. This article studies the intelligent aspect of filtering algorithms and describe a multi‐threshold adaptive filter (MTA filter) for solving this problem. the MTA filter uses a generalized gradient function and a local variance function, which provides the local contextual information as evidence to determine the nature of the filtering for each local neighborhood. A knowledge‐based presegmentation procedure is presented. It applies a threshold operation to extract the local evidence. A belief function is used to combine different evidence and to determine the local filtering strategies. In this way, several simple filters can be combined to form a more efficient and more flexible context dependent filter. As a result, specific filtering is only applied to the region for which it is suitable. Thus, a balanced texture preserving and noise removal effect can be simultaneously achieved.