Cad system for identification of microcalcifications in digitized mammography applying GRNN neural networks

In this paper a CAD system for automatc identification of microcalifications in digitized mamtnography is prcsetitetl. It siiiergicaUy cotribiries tmiltiscale image processing and Artificial Neural Networks (ANNs). Generalid Regression Neural Networks (GHNN) have ben applied in feature selection and identification steps. A comparison Feed Forward Neural Network (FFNN), that are commonly applied in the the downtown step, have been performed. Similar performance has been achieved without the local minimum problems and with a faster training.