Automatic segmentation of whole-slide H&E stained breast histopathology images using a deep convolutional neural network architecture

作者：

Highlights：

• A new framework for segmentation of histopathological images in breast cancer.

• The feasibility of the proposed DCNN-based method is assessed.

• A new tile-wise processing strategy is proposed using dense CRF.

• The method is effective regardless the texture features usual in malignant tumours.

• The framework is available online and can be used as a CAD tool by pathologists.

摘要

•A new framework for segmentation of histopathological images in breast cancer.•The feasibility of the proposed DCNN-based method is assessed.•A new tile-wise processing strategy is proposed using dense CRF.•The method is effective regardless the texture features usual in malignant tumours.•The framework is available online and can be used as a CAD tool by pathologists.

论文关键词：Breast cancer,Segmentation,Deep learning,H&E staining,Whole-Slide Imaging

论文评审过程：Received 31 August 2019, Revised 25 February 2020, Accepted 13 March 2020, Available online 18 March 2020, Version of Record 5 April 2020.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.113387