Abstract:
Breast cancer is the most common cancer in women worldwide, in 2012; breast cancer caused “522 000" deaths, in women worldwide.The conventional visual-inspection method for early breast-cancer detection is impractical, non reproducible, and may bestows ambiguous results. Mammography is an effective technology that has demonstrated the ability to detect breast-cancer at early stages. Early detection of breast-cancer greatly improves the chance of survival. Therefore, a fully automated, accurate, and reliable computerized method it highly needed. This study introduces and proposes two methods for breast-tissue discrimination and early detection of abnormal region on mammogram. In the first one Shock filter has been designed and implemented to play key-role as a fully-automatic technique for enhancing image contrast and help for visual analysis the results showed that this Shock filter has validity as competitive results quality-wise; the second is a semi-automatic technique, which is consist of two phases; firstly based-on A novel Logical algorithm; texture-features were extracted from sub-image as Region-Of-Interest, to be as input to the classifier, to classify the selected breast-tissues into; Fate, Glandular, Dense, Begin, , or cancer. The classifier is based-on Linear Discriminant Analysis (LDA).
The proposed method was evaluated on 250 Sub-Images from mini-MIAS database, and the experimental results have shown that the proposed system achieves accuracy of 96.8%.
