Image Segmentation using Fuzzy C means Clustering with Mahalanobis Distance Norm

Volume 5 Issue 4 October - December 2017

Research Paper

Image Segmentation using Fuzzy C means Clustering with Mahalanobis Distance Norm

Jincy V. Raj *, S. Jini Mol **, Jisha G. Das***, R.S. Sajitha ****

*-**** B.E Scholars, Department of Electronics and Communication Engineering, Bethlahem Institute of Engineering, Karungal, India.

Raj, J, V., Mol, J, S., Das, J, G., and Sajitha, R, S. (2017). Image Segmentation Using Fuzzy C means Clustering With Mahalanobis Distance Norm. i-manager's Journal on Digital Signal Processing, 5(4), 1-9. https://doi.org/10.26634/jdp.5.4.14560

Abstract

In order to map the image, color intensity of the image, or for detecting the object image segmentation is used. It is one of the important procedures used by many of the algorithms. Fuzzy C Means algorithm is one of the effective and powerful image segmentation algorithms compared to all other segments. To describe or explain the dissimilarity in-between Clustered Prototype and the data acquired, FCM uses Euclidean distance to resolve (Zhao et al., 2015). Since the mean information of the cluster is only characterized by the Euclidean distance, both the cluster divergence and noise is made sensitive. Mahalanobis distance is more accurate than the Euclidean distance as a dissimilarity measure when they are used for image segmentation, and they also used to define the mean and covariance of a cluster. The final experimental results show that the Mahalanobis distance is more accurate than the Euclidean distance.