Automated computerized analysis of lesions as a tool to support oncology diagnosis



Cancer is a leading cause of death worldwide. According to the World Health Organizationís cancer report, in 2008, 7.6 million deaths worldwide were caused by cancerous diseases, 0.7 million of those caused by liver cancer and 0.5 million by female breast cancer. For this reason, it is important to diagnose the disease as early as possible and monitor it carefully as it is being treated. Our research aims to develop tools to support the Radiologists and Oncologists diagnosis. It is focused on the development of computational methodologies for disease diagnosis - lesions detection, characterization and monitoring - with specific focus on the computerized analysis of breast lesions in mammograms and liver lesions in 3D multi-phase CT image data. The computational methodologies are based on machine learning tools, such as Bag-of-Visual-Words (BoVW) model, support vector machine (SVM) and deep learning. The breast lesions algorithms are validated using mammography public datasets. The liver cancer research is conducted in collaboration with the CT Abdomen Unit of the Radiology Dept. Sheba Medical center. The new developed tools are evaluated and validated using annotated dataset achieved from this collaboration.



horizontal rule

For problems or questions regarding this web page contact
Last updated: 01/07/08.