Comparative Study of Classification of Histopathological Images
Shraddha Babasaheb Kote, Sonali Agarwal, Ashwini Kodipalli, Roshan Joy Martis
Abstract
The majority of women who suffer from cancer are diagnosed with breast cancer. A type of breast cancer that accounts for about 80% of all other forms of breast cancer is Invasive Ductal Carcinoma (IDC). It is very difficult to diagnose the disease because of its invasiveness. Identification and classification of cancer are of great importance and automated approaches help make efficient usage of time and reduce errors. In this paper, the methods used for classifying histopathological images into Invasive Ductal Carcinoma or non-Invasive Ductal Carcinoma images include standard architectures of Convolutional Neural Networks and machine learning algorithms. The comparative study of the models is performed and is inferred that ResNet50 on the classification produces greater accuracy when compared to other models.