Image Processing and Computer-Aided Detection in Mammographic Images. Annual Report

摘要

Our prior research efforts were directed primarily towards enhancement of and automated detection in side-scan sonar images. The problems associated with these images are very similar to those with mammograms: small, low-contrast targets imbedded in high clutter but a critical requirement to avoid false positives. Our goal is to apply techniques developed for naval images to mammographic images thus making a contribution to the development of a computer-aided diagnosis capability for mammographic data. Second we want to expand the range of image processing techniques applied to naval image data by exploring and modifying methods used by medical researchers. In the process we hope to gain visibility and make contacts in the medical image processing community thus making us privy to cutting-edge methods being developed by medical researchers. Finally, in order to make our work easily accessible to other researchers we will develop and maintain an image analysis toolset with an easy to use interface that can accommodate both naval and medical images.