Features Selection in Mammograms Using Fuzzy Approach

Feature selection plays a key role in identifying and diagnosing breast cancer using computer vision in the mammography which affects the design performance of classifier. This paper introduces for the first time a new fuzzy feature selection approach which utilises fuzzy curve and fuzzy surface to select features from mammogram images. The approach utilizes fuzzy curve to isolate a small set of significant features from the original features according to their significance and eliminate unwanted features. The fuzzy surface is used to eliminate the features dependent on the significant features. This approach reduces the dimensionality of the feature space thus paving a way for a simplified classification scheme for practical application. The result indicates very promising features selected by this approach.