Multi-class Breast Tumor Region Detection for Gigapixel Pathology Images Using Deep Neural Network with Rescale Approach

摘要

Breast cancer has become a worldwide disease in recent years. However, despite its growing prominence, the numberof pathologists equipped to handle these cases is insufficient. Computer-aided diagnosis (CAD) system contributes toreduce costs and improve efficiency of this process. A framework based on convolutional neural networks (CNNs)which could be used to automatically detect the multi-class cancer areas on gigapixel pathology slide images wasproposed. Moreover, combining the slide image characters, rescale and careful data augmentation methods were used totrain the patch-based model with a small dataset. To validate the developed framework, we conducted experiments withBreast Cancer Histology Challenge (BACH) dataset and obtained International Conference on Image Analysis andRecognition (ICIAR) score of 0.582, outperforming the second-place finisher in BACH2018, for the 4-class tissuesegmentation task.