Ταξινόμηση μαστογραφιών για καρκίνο του μαστού με τη χρήση στοιβαγμένων αποθορυβοποιητικών αυτοκωδικοποιητών
Breast cancer classification of mammographies using stacked denoising autoencoders
SubjectΜαστός -- Καρκίνος
Breast cancer is one of the most common cancers among women, and has also become a major cause of death. Medical image analysis is one of the less studied and challenging areas of computer vision and artificial intelligence in general. When we apply machine learning, obtaining reliable results is much more difficult for datasets with abnormalities, such as a dataset of mammographies, than for common applications, since the tissue types and the shape of the organs vary widely from person to person. However, with the use of deep learning, and, specifically, stacked denoising autoencoders, it is possible, with the appropriate tools and parameterization, to have promising, real-world results, that can successfully lead to the correct classification of medical images, hence providing great support to the medical for the early detection of breast cancer.