Adaptive Semi Supervised Support Vector Machine Semi Supervised Learning with Features Cooperation for Breast Cancer Classification

The computer-aided diagnosis (CAD) of breast cancer is becoming increasingly a necessity given the exponential growth of performed mammograms. In particular, the breast mass diagnosis and classification arouse nowadays a great interest. Breast cancer is the most common type of cancer and the second leading cause of cancer deaths (after lung cancer) in women and survival rates critically depend on detection in the initial stages. CAD systems are based on three main steps: segmentation, feature extraction and then classification in order to have a final decision. CAD systems are usually characterized by the large volume of the acquired data that must be labelled in a specific way. However, it is not easy to collect labeled patient records. It takes at least 5 years to label a patient record as 'survived' or 'not survived' What leads to a major problem which is the necessity of an expert to make the labelling operation. That is why the community of statistical learning has attempted to respond to these practical needs by introducing the Semi-Supervised Learning (SSL). For this reason, we have proposed a computer assisted detection system for the diagnosis of this disease based on a particular way on the use of the semi-supervised learning technique using S3VM (Semi Supervised Support Vector Machine) with these different kernel functions. We have made several empirical tests to adapt the parameters of S3VM classifier for better interpretation of mammogram images by changing in each iteration the proportion of labelled data during the training stage and we opted for three different proportions (rarely, low and moderately). Experiments validated DDSM (Digital Database for Screening Mammography) dataset are very encouraging.