BCHisto-Net: Breast histopathological image classification by global and local feature aggregation

作者：

Highlights：

• A novel CNN architecture for 100x breast histopathological image classification.

• A new pre-processing technique to incorporate global and local features is proposed.

• The impact of stain normalization on proposed model is examined.

• The effect of class imbalance problem on the proposed model is analyzed.

摘要

•A novel CNN architecture for 100x breast histopathological image classification.•A new pre-processing technique to incorporate global and local features is proposed.•The impact of stain normalization on proposed model is examined.•The effect of class imbalance problem on the proposed model is analyzed.

论文关键词：Deep learning,Image classification,Convolution neural network,Breast cancer,Histopathological images,Computer - aided diagnosis,Image analysis

论文评审过程：Received 19 March 2021, Revised 15 September 2021, Accepted 5 October 2021, Available online 12 October 2021, Version of Record 16 October 2021.

论文官网地址：https://doi.org/10.1016/j.artmed.2021.102191