Detecting lesions in a mammogram

Computer-aided diagnostics has a growing importance in many medical applications, such as mammography. The main goal of computer-aided tumour-detection is to help radiologists by selecting only the suspected abnormal cases. Identifying abnormal tissues in a mammogram is a great challenge, where conventional image processing methods usually fail. Therefore special methods are needed, which combine the advantages of already known and newly developed image processing algorithms. In this work the possibilities of lesion detection are examined by first extracting the lesion from the background, then characterizing it on the basis of its shape. The achieved results are presented, respectively.