MBTFCN: A novel modular fully convolutional network for MRI brain tumor multi-classification

作者：

Highlights：

• A novel modular fully CNN is proposed for brain tumor MRI classification.

• A residual strip pooling attention is proposed to capture discriminative features.

• Atrous spatial pyramid pooling is adopted to capture tumor contextual features.

• A new classification module is utilized to preserve tumor spatial information.

• The proposed method surpasses other SOTA methods using four public datasets.

摘要

•A novel modular fully CNN is proposed for brain tumor MRI classification.•A residual strip pooling attention is proposed to capture discriminative features.•Atrous spatial pyramid pooling is adopted to capture tumor contextual features.•A new classification module is utilized to preserve tumor spatial information.•The proposed method surpasses other SOTA methods using four public datasets.

论文关键词：Residual Strip pooling Attention,ASPP,Residual network,MRI images,Brain tumors

论文评审过程：Received 13 July 2022, Revised 20 August 2022, Accepted 3 September 2022, Available online 9 September 2022, Version of Record 18 September 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.118776