Abstract: The most common and first method used to detect breast lesions is mammography. Due to less visibility, poor contrast and noisy nature of mammogram it is necessary to detect the small and non-noticeable cancers in early stage. In order to interpret the changes caused due to mental disorder of the breast the contrast of the images are improved while screening the mammograms. The architectural distortion is found in interpreting breast cancers as well as microcalcification and mass on mammograms. However, it is more difficult to detect architectural distortion than microcalcification and mass. The original mammogram image is decomposed using wavelet decomposition and gabor features are extracted from the original image Region of Interest (ROI). The ability of these features in detecting microcalcification is done using Backpropagation Neural Network (BPNN). The proposed classification approach is applied to a database of 322 dense mammographic images, originating from the MIAS database.This paper presents various techniques used for automatic enhancement and segmentation of microcalcifications in mammographic images. These techniques consists of three different stages which includes preprocessing stage, feature extraction stage and classification stage. The paper represents the proposed system in which these stages can be implemented using histogram equalization, contrast stretching and image segmentation.

Keywords: Mammography, Microcalcifications, Preprocessing, Feature Extraction, Classification, Histogram Equalization, Contrast Stretching, Backpropagation Neural Network and Image Segmentation.