Automatic Diagnosis and Classification of Abnormalities in Digital X-ray Mammograms

Mammography remains the most effective tool for the early detection of breast cancer and ComputerAided Diagnosis (CADx) is usually used as a second opinion by the radiologists. The main objective of our study is to introduce a method to generate and select the features of suspicious lesions in mammograms and classifying them by using support vector machine, in order to build a CADx system to discriminate between malignant and benign parenchyma. Our method has been verified with the well-known Mammographic Image Analysis Society (MIAS) database and we have used the Receiver Operating Characteristics (ROC) to measure the performance of our method. The experimental results show that our method achieved an overall classification accuracy of 96.36%, with 96.77% sensitivity and 95.83% specificity in the training phase and achieved an overall classification accuracy of 94.29%, with 94.11% sensitivity and 94.44% specificity in the testing phase. Key–Words: Mammography, Breast, Computer Aided Diagnosis, Support Vector Machine, ROC analysis.