Weakly-supervised Learning Using Pretraining for Classification in HER2 Immunohistochemistry Image of Breast Cancer
Zhengnan Wang, Yeting Ma, Yali Zheng, Peiqin Feng, Fangbo Yu · 2021
Recently supervised deep learning method has achieved good performance in image classification tasks. However, it is very difficult to annotate pathological images accurately for supervised learning tasks. So, the limited amount of labeled data brings great challenges to the supervised learning model. In this paper we propose a weakly-supervised learning method which combines the pretraining technology of transfer learning with deep learning in the HER2 immunohistochemistry (IHC) pathological image classification task of breast cancer. It is worth mentioning that on the network architecture of VGG16 model, we train the model with three different images from pathological images, apply the pretraining model to the HER2 IHC image classification task of breast cancer. The experimental results show that the weakly-supervised learning implemented by the pretraining technology of transfer learning can significantly improve the performance of HER2 IHC pathological image classification task.