Towards a neural network based system for skin cancer diagnosis

This paper reports an investigation into the application of a multilayer perceptron to the diagnosis of skin melanoma. The lesions are classified as either benign or malignant based on information relating to the shape of their outline. The results obtained by the standard back-propagation learning algorithm are compared to those attained by various dynamic network design strategies. These demonstrate that the results attained with the standard multilayer perceptron can be improved upon by modifying the network architecture during the training process.<
>