This paper presents a high accuracy computer-aided system to detect microcalcifications and classify them into benign or malignant. The microcalcifications detection procedure is mainly based on a combination of adaptive histogram equalization, median filtering, morphological operations and thresholding. The contribution of this type of decomposition is denoising and enhancing regions of interests (ROI) containing microcalcifications. Feature extraction is performed on detected microcalcifications and their surrounding tissues using the most well known features in the literature then feature selection is done to reduce these features to the most important features according to the accuracy. SVM and K-NN (conventional, fuzzy and voting) classifiers are used. Our results show that the developed methods are effective for quantifying the classification of benign and malignant microcalcifications with an accuracy of 100%.