Auxiliary System For Computer-Aided Evaluation of Breast Calcifications Based on Digital Image Processing and Artificial Neural Networks

Early diagnosis still represents the best approach in the prevention and control of breast cancer, the second most frequent form of cancer worldwide. In this context, mammography has been largely used as a major method for disease early detection, as it aids the early identification of calcification clusters that can be or can later become tumors. This paper discusses the use of Wavelet Transforms to help highlight calcification areas in combination with digital image morphological techniques for feature extraction from regions of interest in digital mammography. The results obtained in this work are being upgraded to a next step, which will use a MultiLayer Perceptron artificial neural network to classify calcifications according to the most applicable category in the Breast Imaging Reporting and Data System (BIRADS).