Algorism of extracting nuclear and membrane for diagnosing early breast cancer

As technology of image diagnosis develops, the number of early breast cancer patients tends to increase, today. Then, in order to improve their QOL, we should select a cure for them. Breast cancer is classified as four disease types by expression of three proteins (ER, PgR, HER2). And, ER and PgR express in nuclear and HER2 expresses in membranes. So, it is important that we extract nuclear and membrane exactly. Then, we propose the methods of extracting nuclear by watershed algorithm and using membrane, and membrane by using canny filter. By using our methods, we can extract nuclear and membrane more exactly than by general method.