Classification of breast lesions using artificial neural network

This paper presents a study on classification of breast lesions using artificial neural networks. Thirteen morphological features have been extracted from breast lesion cells and used as the neural network inputs for the classification. Multilayered Perceptron, Radial Basis Function and Hybrid Multilayered Perceptron networks were used to perform the classification task. Unlike the previous studies that only classify the lesion into benign and malignant, this study extends the breast lesions classification into four categories that are malignant, fibroadenoma, fibrocystic disease and other benign cells. The three neural networks were trained and compared using 1300 data samples. The classification results indicating that all the networks give good overall diagnostic performance. However, only Hybrid Multilayered Network that provides 100% accuracy, sensitivity and specificity.