Mammographic masses classification: novel and simple signal analysis method

Breast cancer represents the leading cause of fatality among cancers for women and there is still no known way of preventing this pathology. Computer-aided analysis systems could be very helpful to improve both the sensitivity and the specificity. Presented is a computer-aided diagnosis system for malignant/benign masses classification of regions of interest in mammograms based on wavelet transform decomposition and an artificial neural network classifier is shown. The main novelty of the system is to consider only small 1D signals crossing the region of interest, allowing one to drastically reduce the amount of data to be processed. An experimental analysis performed on a set of images from the DDSM database has shown the effectiveness of the proposed method.