Date of Award
2015
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Electrical and Computer Engineering
Committee Chair
Reza R. Adhami
Committee Member
John Jarem
Committee Member
David Pan
Committee Member
Sivaguru Ravindran
Committee Member
Seong-Moo Yoo
Subject(s)
Diagnostic imaging--Digital techniques, Breast--Cancer--Imaging, Breast--Radiography
Abstract
Screening mammography often incorporates a computer aided diagnosis (CAD) scheme in its procedure to increase the accuracy of detecting gradual changes in breast tissues. One method for detecting gradual changes in temporal mammograms is through registration algorithms. Most registration algorithms require segmented mammograms as their inputs. The performance of registration algorithms and, hence, the performance of the CADs, are directly proportional to the quality of the segmented mammograms. Segmented mammograms include breast landmarks such as the nipple, the breast boundary, and the pectoral muscle. In this research, we designed, implemented, and evaluated a new segmentation algorithm for pectoral muscle detection. The presented algorithm uses global thresholding to locate two lines that represent the pectoral muscle. The final edge of the detected pectoral muscle is curved using the sigmoid function. The detected pectoral muscle is removed from mammograms with mediolateral oblique (MLO) view and applied to a registration algorithm. An iterative registration algorithm that uses structural similarity (SSIM) index is developed to compute the optimal transformation that maps information in a temporal mammogram pair. The performance of the SSIM algorithm is compared with those of the correlation (CORR) coefficient and mutual information (MI) algorithms. It is shown that the SSIM outperforms the CORR and MI in terms of error rate.
Recommended Citation
Al-Ghaib, Huda, "Temporal mammogram segmentation for improving early breast cancer detection" (2015). Dissertations. 72.
https://louis.uah.edu/uah-dissertations/72