Reason: Access restricted by the author. A copy can be requested for private research and study by contacting your institution's library service. This copy cannot be republished
A computer aided detection sytem for the evaluation of breast cancer
thesisposted on 2017-02-27, 00:28 authored by Sujit, Sheeba Jenifer
For many years, breast cancer has been one of the leading causes of death among women. Early detection of breast cancer is the key to improve the survival rate. Mammography is currently the most efficient tool for early detection of breast cancer. A large number of digital mammograms generated each year needs accurate and fast interpretation of images. The drawback of mammography is that the digital mammograms are difficult images to be interpreted visually. Non-cancerous lesions can be misinterpreted as a cancer (false-positive value), while cancers may be missed (false-negative value). A Computer-Aided Detection (CAD) system can help radiologists in this difficult task of interpreting digital mammograms. Though there are many CAD techniques developed today, clinical success depends on CAD having a high sensitivity, specificity and accuracy and the reader taking appropriate action when interpreting the CAD prompts. This balance is not easy to achieve. The motivation behind this research work is to develop a CAD system with high accuracy, sensitivity and specificity in the range of 95% to 100%. The work plan of the proposed research consists of four stages: The first stage involves the development of pre-processing methods for contrast enhancement of digital mammograms.The second stage of the proposed CAD is the development of suitable algorithms for segmentation of regions of interest using Otsu algorithm and data clustering techniques. The third stage involves the selection and extraction of image features tailored for the analysis of digital mammograms. Thirteen textural features for four spatial orientations: 0°, 45°, 90° and 135° are extracted from Gray-level co-occurrence matrices (GLCM) for the purpose of image analysis. The last stage is the evaluation of Neural network architectures and Support Vector Machines (SVM) aimed at using CAD system as a decision making aid for automatic detection of breast cancer. The Mammographic Image Analysis Society (MIAS) database of digital mammograms is used to evaluate the effectiveness of the developed CAD system. The overall performance of the developed system is compared with previously established methods based on the performance measures such as accuracy, specificity and sensitivity.