Open AccessComparative analysis of modern automated algorithms image segmentation
Author(s) -
O.M. Lisenko,
A.YU. Varfolomєєv
Publication year - 2012
Publication title -
èlektronika i svâzʹ
Language(s) - English
Resource type - Journals
eISSN - 2312-1807
pISSN - 1811-4512
DOI - 10.20535/2312-1807.2011.16.5.247555
Subject(s) - mean shift , cluster analysis , segmentation , image segmentation , artificial intelligence , segmentation based object categorization , computer science , scale space segmentation , pattern recognition (psychology) , graph , expectation–maximization algorithm , computer vision , image (mathematics) , algorithm , mathematics , maximum likelihood , statistics , theoretical computer science
Unsupervised image segmentation algorithms based on-mean clustering, expectation-maximization, mean-shift, normalized graph cut, weighted aggregation, statistical region merging, JSEG, HGS and ROI-SEG are considered. The results of segmentation obtained by mentioned algorithms on textural, satellite and natural images are presented. The analysis of quality and segmentation speed of each algorithm realization is performed