Record Details

HoEnTOA: Holoentropy and Taylor Assisted Optimization based Novel Image Quality Enhancement Algorithm for Multi-Focus Image Fusion

NOPR - NISCAIR Online Periodicals Repository

View Archive Info
 
 
Field Value
 
Title HoEnTOA: Holoentropy and Taylor Assisted Optimization based Novel Image Quality Enhancement Algorithm for Multi-Focus Image Fusion
 
Creator Singh, Vineeta
Kaushik, Vandana Dixit
 
Subject Contourlet transform
Medical imaging
Optimization based fusion technique
Taylor series
Two level image fusion
 
Description 875-886
In machine vision as well as image processing applications, multi-focus image fusion strategy carries a prominent
exposure. Normally, image fusion is a method of merging of information extracted out of two or more than two source
images fused to produce a solitary image, which is much more instructive as well as much suitable for computer processing
and visual perception. In this research paper authors have devised a novel image quality enhancement algorithm by fusing
multi-focus images, in short, termed as HoEnTOA. Initially, contourlet transform is incorporated to both of the input images
for generation of four respective sub-bands of each of input image. After converting into sub-bands further holoentropy
along with proposed HoEnTOA is introduced to fuse multi-focus images. Here, the developed HoEnTOA is integration
of Taylor series with ASSCA. After fusion, the inverse contourlet transform is incorporated for obtaining last fused image.
Thus, the proposed HoEnTOA effectively performs the image fusion and has demonstrated better performance utilizing
the five metrics i.e. Root Mean Square Error with a minimum value of 3.687, highest universal quality index value of
0.984, maximum Peak Signal to Noise Ratio of 42.08dB, maximal structural similarity index measurement of 0.943, as well
as maximum mutual information of 1.651.
 
Date 2021-10-05T11:50:07Z
2021-10-05T11:50:07Z
2021-10
 
Type Article
 
Identifier 0975-1084 (Online); 0022-4456 (Print)
http://nopr.niscair.res.in/handle/123456789/58230
 
Language en
 
Publisher NIScPR-CSIR, India
 
Source JSIR Vol.80(10) [October 2021]