DocumentCode :
701299
Title :
A non-iterative approach to initial region estimation applied to color image segmentation
Author :
Portillo-Garcia, Javier ; Alberola-Lopez, Carlos ; Tardon-Garcia, Lorenzo Jose ; Trueba-Santander, Juan Ignacio
Author_Institution :
SSR-ETSI Telecomunicación-UPM, Ciudad Universitaria s/n. 28040 Madrid, Spain
fYear :
1996
fDate :
10-13 Sept. 1996
Firstpage :
1
Lastpage :
4
Abstract :
A non-iterative segmentation approach is developed to generate a fast initial estimation of the layout of the different color textures presented in the original image mainly based on hypothesis testing. Most of the proposed methods have a tremendous computational burden which make them difficult to be implemented in a real-time working processor. We will compare our method with a known iterative clustering algorithm that guides to similar results with much higher computational cost. We present two examples that show similar results and compare the computational cost for each case. Spotty resemblance caused by pixel oriented decision is diminished in both cases by modeling regions as Markov Random Fields.
fLanguage :
English
Publisher :
ieee
Conference_Titel :
European Signal Processing Conference, 1996. EUSIPCO 1996. 8th
Conference_Location :
Trieste, Italy
Print_ISBN :
978-888-6179-83-6
Type :
conf
Filename :
7083024
Link To Document :
بازگشت