Segmentation of mammographic images for computer aided diagnosis

Computer-aided diagnosis systems are currently at the heart of many clinical protocols. This book puts forward a hierarchical architecture for the design of a robust and efficient CAD tool for breast cancer detection. It focuses on the reduction of false alarms rate through the identification of image regions of foremost interest (potential cancerous areas). The dynamic range of the image is stretched to enhance the contrast between tissues and background and favors accurate breast region extraction. Then follow pectoral muscle segmentation since it regularly tampers breast tissue analysis. Extracting pectoral muscle tissues is both hard and challenging due to its overlap with dense tissues. To overcome this difficulty, a validation process followed by a refinement strategy is proposed to detect and correct the segmentation imperfections. In the last chapter dealing with breast density analysis, to address the inter-variability in gray levels distributions, an optimized gray level transport map is introduced for contrast standardization. With this technique, dense region areas computed using simple thresholding are highly correlated to density classes from an annotated dataset.