Robust Gaussian-base radial kernel fuzzy clustering algorithm for image segmentation

To perform the image segmentation task, in this Letter, a kernel fuzzy C-means algorithm is introduced, strengthened by a robust Gaussian radial basis function kernel based on M-estimators. It is well-known that these kernels consider the squared difference as a similarity measure, which is not robu...

Full description

Saved in:
Bibliographic Details
Published in:Electronics letters Vol. 55; no. 15; pp. 835 - 837
Main Authors: Mújica-Vargas, Dante, Carvajal-Gámez, Blanca, Ochoa, Genaro, Rubio, José
Format: Journal Article
Language:English
Published: The Institution of Engineering and Technology 25.07.2019
Subjects:
colour images
fuzzy set theory
Gaussian processes
Image and vision processing and display technology
image segmentation
image segmentation task
M‐estimators
pattern clustering
radial basis function networks
robust Gaussian radial basis function kernel
robust Gaussian‐base radial kernel fuzzy clustering algorithm
robust Gaussian radial basis function kernel
image segmentation task
pattern clustering
image segmentation
fuzzy set theory
Gaussian processes
radial basis function networks
colour images
robust Gaussian-base radial kernel fuzzy clustering algorithm
M-estimators
ISSN:0013-5194, 1350-911X, 1350-911X
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:To perform the image segmentation task, in this Letter, a kernel fuzzy C-means algorithm is introduced, strengthened by a robust Gaussian radial basis function kernel based on M-estimators. It is well-known that these kernels consider the squared difference as a similarity measure, which is not robust to atypical data. In this regard, the main motivation of this contribution is to improve the atypical information tolerance of these kernels, in order to make a better clustering of pixels. Experimental tests were developed considering colour images. The robustness and effectiveness of this proposal are verified by quantitative and qualitative results.
ISSN:0013-5194
1350-911X
1350-911X
DOI:10.1049/el.2019.1281