Off-line mammography screening system embedded with hierarchically-coarse-to-fine techniques for the detection and segmentation of clustered microcalcifications

An Off-line mammography screening system is used in pre-screening mammogram s to separate high-risk mammograms From most normal cases. Off-lint system can run before radiologist's review and is particularly useful in tl-le nat ional breast cancer screening program which usually consists of high percen tage of normal cases. Until now, the shortcomings of online detection of cl ustered microcalcifications from a mammogram remain in the necessity of man ual selection of regions of interest. The developed technique focuses on de tection of microcalcifications within a region of interest indicated by the radiologist. Therefore, this kind of system is not efficient enough to pro cess hundreds of mammograms in a short time without a large number of radio logists. In this paper, based on a "hierarchically-coarse-to-fine" approach , an off-line mammography screening system for the detection and segmentati on of clustered microcalcifications is presented. A serial off-line procedu res without any human intervention should consider the complexity of organi zation of mammograms. In practice, it is impossible to use one technique to obtain clustered microcalcifications without consideration of background t ext and noises from image acquisition, the position of breast area and regi ons of interest. "Hierarchically-coarse-to- fine" approach is a serial proc edures without any manual operations to reduce the potential areas of clust ered microcalcifications from a mammogram until clustered microcalcificatio ns are found. The reduction of potential areas starts with a mammogram, thr ough identification of the breast area, identification of the suspicious ar eas of clustered microcalcifications, and finally segmentation of clustered microcalcifications. It is achieved hierarchically from coarse level to fi ne level. In detail, the proposed system includes breast area separation, e nhancement, detection and localization of suspicious areas, segmentation of microcalcifications, and target selection of microcalcifications. The syst em separates its functions into hierarchical steps and follows the rule of thumb "coarse detection followed by fine segmentation" in performing each s tep of processing. The decomposed hierarchical steps are as follows: The sy stem first extracts the breast region from which suspicious areas are detec ted. Then precise clustered microcalcification regions are segmented from t he suspicious areas. For each step of operation, techniques for rough detec tion are first applied followed by a fine segmentation to accurately detect the boundaries of the target, regions. With this "hierarchically- coarse-t o-fine" approach, a complicated work such as the detection of clustered mic rocalcifications call be divided and conquered. The effectiveness of the sy stem is evaluated by three experienced radiologists using two mammogram dat abases from the Nijmegen University Hospital and the Taichung Veterans Gene ral Hospital. Results indicate that the system can precisely extract the cl ustered microcalcifications without human intervention, and its performance is competitive with that of experienced radiologists, showing the system a s a promising asset to radiologists.