Histopathology Grading Identification of Breast Cancer Based on Texture Classification Using GLCM and Neural Network Method

Abstract

Breast cancer is the leading type of malignant tumor which is observed in women. The effective treatment depends on its early diagnosis. The gold standard of breast cancer is histopathologic examination of cancer cells. The determination of the grading in breast cancer is determined by three factors: pleomorphic, tubular formation and cell mitosis. This paper uses pleumorfic and tubular formation pattern from breast cell histopathology images. The proposed system consists of four major steps : preprocessing, segmentation, feature extraction and classification. We use k - means clustering method for image segmentation and use Gray level Cooccurence Matrix (GLCM) for feature extraction with four features (i.e. angular second moment, contrast feature, entropy feature, and variance feature). The final step is grading classification which uses Backpropagation Neural Network. Some of important parameters will be variated in this process such as learning rate and the number of node in hidden layer. The research gives good result for the identification of breast cancer grading with 88% accuracy, 85% sensitivity, and 80% specificity.