Comparison of Distance Measures on Fuzzyc-means Algorithm for Image Classification Problem

A study on the use of two different distance measures, Euclidean distance and divergence distance, for FCM is conducted for an image classification problem in this paper.Conventional FCM algorithm which uses Euclidean distance measure utilizes only mean information from an image block for its feature while FCM algorithm with divergence utilizes both of variance and mean information. Evaluationson a set of Caltech databaseshow that Divergence-based FCM gives higher accuracy when compared with some conventional algorithms with Euclidean distance.