The Improved Membership Matrix Initialization Method of FCM for Image Segmentation
2014 ◽
Vol 989-994
◽
pp. 3743-3746
Keyword(s):
The performance of FCM for image segmentation directly subjects to the initialized membership matrix. This paper proposed twice FCM method to solve the membership matrix initiation problem. The image is spared to a blurred image at first, and then uses the FCM for the blurred image to obtain an iterative result, in which the membership matrix is taken as the initialized membership function of the FCM for the original image processing. This method overcomes the random membership initialization method cannot convergence to the optimum point of the objective function of FCM for the image segmentation at some extend, furthermore, it can obtain better results than the traditional FCM method.
2014 ◽
Vol 1070-1072
◽
pp. 2041-2044
◽
2021 ◽
Vol 12
(5)
◽
pp. 862-865
Keyword(s):
2015 ◽
Vol XL-7/W3
◽
pp. 703-706
2017 ◽
Vol 32
(04)
◽
pp. 1850012
◽
2014 ◽
Vol 945-949
◽
pp. 1899-1902
Keyword(s):
2014 ◽
Vol 496-500
◽
pp. 1834-1839
Keyword(s):
2014 ◽
Vol 1
(2)
◽
pp. 62-74
◽
Keyword(s):
2013 ◽
Vol 734-737
◽
pp. 2912-2916