Artificial neural classification of clustered microcalcifications on digitized mammograms

Investigates the potential utility of artificial neural networks as a decision-making aid to radiologists in the analysis of mammograms. Multi-layer neural networks with a backpropagation algorithm were trained for differentiating malignant from benign microcalcification clusters on the basis of both the clinical parameter "age" and radiological features extracted from digitized mammographic images. Radiological features associated with clustered microcalcifications were extracted from digitized images using mathematical morphology and image analysis. These parameters were analyzed using statistical analysis. The most relevant shape parameters were then given as input to the artificial neural network. Successful experimental results obtained by a three-layer neural network indicate that such an approach may provide a potentially useful tool to radiologists in the clustered microcalcification classification task.