Automatic Prediction of Breast Cancer Metastasis Stages via Deep Convolutional Networks

Automatic and accurate detection of breast cancer metastases is meaningful for reducing the workload of the pathologists and reducing the cost of diagnosis. In this paper, we presented a learning-based method for automated prediction of breast cancer metastasis stages. This method is trained and validated on Camelyon17 challenge datasets. The method consists of three parts: tissue extraction, tumor region detection and cancer metastasis stage prediction. Index Terms – Cancer metastases prediction, deep convolutional networks, Camelyon17 challenge