الفهرس | Only 14 pages are availabe for public view |
Abstract The contribution of this thesis is the development of a Computer Aided Diagnosis (CAD) system. CAD system assesses the diagnosis of breast cancer, reduces the error margin in tumor diagnosis and aids in early detection of the disease. This thesis presents an automatic scheme to perform both detection and segmentation of breast masses. Firstly, the breast region (i.e. the region of interest ROI) is determined and extracted from the whole mammogram image. This ROJ is detected using a new method for the identification and removal of the pectoral muscle in Medio-Lateral Oblique (MLO) view of mammogram. This method uses an iterative thresholding technique to separate the muscle. Afterword, the breast border is extracted. This step for determining the ROJ is an essential pre-processing step. Primarily it allows the search for abnormalities to be limited to the region of the breast without undue effect from the background of the mammogram. Secondly, an automatic segmentation algorithm is proposed to perform an accurate identification of the mass region. The proposed algorithm segments the region contains abnormalities from the normal breast tissue using a Region growing method. Region growing and pixel aggregation is applied in the ROI begins as a single pixel and grows based on surrounding pixels that satisfy certain similarity criterion. In this thesis, a new method for the threshold (i.e. the similarity criterion) selection was proposed to enhance the region growing technique. Keywords Mamrnograms, Computer Aided Diagnosis, Pectoral Muscle Suppression, Breast Border Detection, Segmentation of Masses . |