BREAST CANCER CLASSIFICATION USING ARTIFICIAL NEURAL NETWORK AND TRANSFER LEARNING ON HISTOLOGY IMAGES
Gagan Jindal, Geetanjali Babbar · Journal of Critical Reviews · 2020
Abstract From past few years, the breast cancer is the second main cause for deaths of women around the world. Previously, breast cancer cell was being detected with the help of biopsy and pathologist which was the only way. In recent years, computer aided technique machine learning is introduced in which computer first trains on specific patterns on breast cancer cell histopathological images and then detect whether cell is cancerous or not. In this paper, comparison of two approaches is made for classification of breast cancer into benign and malignant using histopathological images. First approach is artificial neural networks in which features are being carried out with the help of 30 hidden units present in three hidden layers (10 units in each layer) and after then features are embed in softmax layer for classification of benign and malignant cancer. In second approach, transfer learning technique VGG19 model is use for classification. The result shows VGG19 transfer learning model perform better in comparison to artificial neural network where VGG19 achieve 98.4% accuracy and artificial neural network achieve 83% accuracy