HER2-MCNN: a HER2 classification method based on multi convolution neural network
Xingang Wang, Cuiling Shao, Jiandong Lv, Liang Hu, Na Li · 2021
Breast cancer is the most common malignant tumor in women and the second leading cause of cancer death. In breast cancer, about 20% to 30% of patients would have the positive phenomenon of human epidermal growth factor receptor 2 (HER2), also known as HER2 positive breast cancer. So far, different models have been proposed by using computer network to help HER2 classification. However, the existing single network model has a single receptive field scale, which can not extract multiple abstract features of the image well. Therefore, we designed a compound network, in which we use different network to extract the image features. Then the image information extracted from different networks was fused by linear weighted fusion algorithm, after that the weight in linear weighted fusion was optimized by experimental comparison. Finally, the SVM classifier was used to output the HER2 image categories. The experimental results showed that the proposed network has high accuracy for HER2 classification.