Performance analysis of Neural Network based classification technique for Mammogram Images

摘要

This paper presents experimental work on mammogram image analysis. Texture analysis is carried out using segmentation technique. Here, statistical method have been used to extract features from the segmented tumour area. The obtained features are classified using different classifiers such as Radial basis function, Main Feed forward and Main Fitnet method. The method was tested on 100 clinical data. The RBF classifier achieved an accuracy of 91% Main fit net accuracy of 99.20% Main Feed forward accuracy about 97%.