Multi-classification of Histopathological Images based on Convolutional Neural Networks
Xiaoting Wang, Jiying Li, YanDong Lu · International Conference on Frontiers of Electronics, Information and Computation Technologies · 2021
Breast cancer is the main cause of harm to women's health, and pathological examination is the key means of diagnosis of breast cancer, so the accurate classification of histopathological images is of great significance in clinical application. The classification of benign (B), malignant (M) and eight subtypes (adenosis (A), fibroadenoma (F), tubular adenoma (TA), phyllodes tumor (PT), ductal carcinoma (DC), lobular carcinoma (LC), mucinous carcinoma (MC) and papillary carcinoma (PC)) of breast cancer was studied. The convolutional neural network was used to realize the automatic feature learning and automatic classification of images. A preprocessing method suitable for HE staining images and a training strategy for global fine-tuning were proposed. The dataset was expanded by the enhancement technology. Comparison experiments were carried out on two different models, Googlenet and Resnet50, and the results were compared with the existing research results. The results show that the proposed method can produce higher classification accuracy.