Diagnostic Screening of Digital Mammograms Using Wavelets and Neural Networks to Extract Structure
Barry L. Kalman · Open Scholarship Institutional Repository (Washington University in St. Louis) · 1998
. As the primary tool for detecting breast carcinoma, mammography provides visual images from which a trained radiologist can identify suspicious areas that suggest the presence of cancer. We describe an approach to image processing that reduces an image to a small number of values based on its structural characteristics using wavelets and neural networks. To illustrate its utility, we apply this methodology to the automatic screening of mammograms for mass lesions. Our results approach performance levels of trained human mammographers. Keywords: structured knowledge extraction, classification of structured information, unsupervised learning of hierarchical structure, medical diagnosis, linear-output sequential recursive auto-associative memory. 1. Introduction Increasingly, modern medicine relies on a vast array of imaging studies for diagnosis. Mammography, in particular, supports efforts to screen for and detect breast carcinoma, a disease that will affect one in nine women over ...