DocumentCode :
3744385
Title :
Overlapped cells separation algorithm based on morphological system using distance minimums in microscopic images
Author :
Vahid Khodadadi;Emad Fatemizadeh;S. Kamaledin Setarehdan
Author_Institution :
Department of biomedical Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran
fYear :
2015
Firstpage :
263
Lastpage :
268
Abstract :
It is important to distinguish overlapped cell for tracking and segmentation biological cells from images. In this research, a novel or comprehensive method is provided with morphological features of cell and minimum distance for overlapped cells separation (OCS). it´s necessary to say that this algorithm is not based on type and number of Overlapped cells. In this presented method based on a distance chart. At first, every cell or rectangle of cell is separated of the original image. Also, multiplex cell is distinguished by using minimum perimeter polygon (MPP) method. Distance chart is achieved on the base of morphological procedure and Computing of cell boundary distance to Cell middle axis in different epochs. By achieving the absolute minimum distance chart, candidate cutting points is detected and separated. Experimental results show the offered method is very effective and the work shows great assistance for the cell Cytoplasm segmentation. At the end, recognition accuracy and 97.4% separation for cells is achieved by this method.
Keywords :
"Image segmentation","Shape","Morphological operations","Transforms","Clustering algorithms","Redundancy","Arrays"
Publisher :
ieee
Conference_Titel :
Biomedical Engineering (ICBME), 2015 22nd Iranian Conference on
Type :
conf
DOI :
10.1109/ICBME.2015.7404153
Filename :
7404153
Link To Document :
بازگشت