Hierarchical Approach for Breast Cancer Histopathology Images Classification

Nidhi Ranjan, Pranav Vinod Machingal, Sunil Sri Datta Jammalmadka, Veena Thenaknidiyoor, and A. D. Dileep

Abstract: Cancer is one of the highest death causing diseases in the world, with breast cancer being the highest in women. Early stage detection of breast cancer can make the treatment process more effective. Typically the identification of stages of a cancer is performed by pathologists by analysing histopathology images/slides.Analysing histopathological images is a highly tedious task that requires long attention span. In this paper we propose an approach to classify histopathology images. In this work we lead to a CNN-based Hierarchical classifier to classify histopathology images into various stages of cancer. Deep learning in the field of Medical Image Analysis has shown propitious results.The effectiveness of the proposed CNN-based Hierarchical classifier is studied using the BACH challenge dataset. The proposed method achieved an accuracy of 95% on our validation set.

Author Affiliation: National Institute of Technology Goa, Indian Institute of Technology Mandi

Keywords: breast-cancer histopathology, hierarchical, classification of breastcancer, deep learning