Description of Invention
The leading cause of cancer-related death in women is breast cancer. Only early detection can save a life but, traditional manual bioscopy is time-consuming. Digital pathology can be crucial in detecting breast cancer by automatic nuclei segmentation and classification using deep learning. We propose a comprehensive analysis of a deep learning-based approach to segment and classify breast cancer nuclei. We aimed to identify the best model for classifying the negative, weak, moderate, and strong nuclei. We have 220 ROI from 40 WSI of ER-IHC. Here we found Overall Recall: 0.85 and Overall F1 Score: 0.89.