Mammogram images classification using Gray Level Co-occurence Matrices

The aim of study was to classify mammogram images using Gray Level Co-occurence Matrices - GLCM. In study, morphological image processing was used as a preprocessing method to detect breast regions. Both GLCM texture features; contrast, correlation, energy and homogeneity and statistical features; mean, variance, skewness and kurtosis are extracted from the breast regions on images. The images are classified by using (Support Vector Machines - SVM) and (Nearest Neighbor - NN) methods. GLCM and statistical features are examined on classification accuracy.