الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 93 |  |  |
| | | from | 93 | | |
Abstract
Breast cancer (BC) is a common disease in which cells in the breast grow out of control, accurate and early detection of breast cancer is very important to reduce the mortality and morbidity. Lately, it has been found a strong relationship between temperature variation and breast cancer .Previous studies expose that thermography is a good tool for early detection of the breast cancer. In this thesis, a new automatic system will be introduced for the early detection of the breast cancer using thermal images and distinguishing between normal and abnormal breasts. The proposed system is based on combining textural features and histogram of oriented gradients and bag of thermal breast images and then classifying those using three different classifiers: (i) Support vector machine; (ii) Decision tree, and k-Nearest Neighbor. This proposed system provides an automatic classification of the breast cancer using image analysis accurately in low elapsed time. Experimental results showed that cubic SVM has a maximum accuracy of 98.9%, a sensitivity of 98.9%, and a specificity of 99%. When comparing the proposed system with the relevant systems, it’s approved to be more accurate with low elapsed time in learning and testing phase that can help the clinicians in the automatic diagnosis of the breast cancer.