Premium
Contour extraction of glomeruli by using genetic algorithm for edge patching
Author(s) -
Zhang Jun,
Hu Jinglu,
Zhu Hong
Publication year - 2011
Publication title -
ieej transactions on electrical and electronic engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.254
H-Index - 30
eISSN - 1931-4981
pISSN - 1931-4973
DOI - 10.1002/tee.20649
Subject(s) - dilation (metric space) , canny edge detector , edge detection , artificial intelligence , genetic algorithm , pattern recognition (psychology) , enhanced data rates for gsm evolution , computer science , operator (biology) , algorithm , image (mathematics) , computer vision , mathematics , image processing , biology , gene , machine learning , biochemistry , repressor , combinatorics , transcription factor
Glomeruli extraction is a vital step in computer‐aided diagnosis systems of kidney disease. Since there are not only glomeruli but also other tissues in an image, when detecting the edges of glomeruli, lot of noises caused by other tissues will be detected at the same time. These noises cause discontinuous edges of glomeruli when some operation, such as labeling, is applied to denoise. According to this characteristic, this article proposes a contour extraction method based on genetic algorithm (GA) for edge patching. First, a Canny operator is applied to obtain the edges of glomeruli with noises. Then labeling and other operations such as dilation, thinning and cross‐point deletion are applied to markedly remove the noises. After the above operations, GA is finally used to search for optimal patching segments to join the discontinuous edges together and a closed curve with highest fitness would be able to form the contour of glomeruli. Experimental results show that the proposed method performs well for the renal biopsy images. © 2011 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.