Boundary Detection of Incomplete Shapes in Breast Thermal Images Using Statistical Shape Modeling
摘要
Breast cancer is a life-threatening disease for women. Early detection of breast cancer improves the patient's chance of survival. Several detecting methods include ultrasonography, mammography, thermography, and biopsy. Thermography is a cheap and accessible detection technique. Thermography images and computer-aided diagnosis algorithms help patients with regular monitoring. To improve the detection and classification accuracy of the lesions, a definition of the breast region of interest is necessary. In this research, we propose a statistical shape model to represent the boundary of the breast in lateral left and right thermal images. This model is later used for the segmentation of the breast. The proposed technique effectively defined the organ's boundary using some manual landmarks. An automatic scheme defined the border of the breast as well. The proposed approach was successful in the segmentation of incomplete shapes, too.