Detection and classification of masses in mammograms using ICA

This work is focused on the detection and classification of masses in mammograms. The masses are one of the main indicators of the existence of breast cancer, malignant or benign. For this, here has been developed a detection and classification system of masses in mammograms. This system consists of a set of artificial intelligence techniques such as feature extraction (Principal Component Analysis, PCA, and Independent Component Analysis, ICA) and pattern classification using neural networks (Multilayer Perceptron) and SVM classifiers. The detection and classification of lesions in mammograms is an extremely hard task because of the large variability that the lesions may have, despite being of the same type. In addition, there is the variability that the mammograms may present depending on the breast tissue predominant in the breast and its distribution. Often, this may cause that cancer might appear obscured, masked by the surrounding breast tissue which makes it very difficult to detect.