CVM-Cervix: A hybrid cervical Pap-smear image classification framework using CNN, visual transformer and multilayer perceptron

作者：

Highlights：

• This study performs a 11-class classification task for cervical cells, and it is the work with the most cell types in existing literature.

• A new method (CVM-Cervix) is proposed, which uses CNN, VT and MLP modules for classification.

• The performance of our CVM-Cervix exceeds that of existing methods and achieves a good result of 91.72% accuracy.

摘要

•This study performs a 11-class classification task for cervical cells, and it is the work with the most cell types in existing literature.•A new method (CVM-Cervix) is proposed, which uses CNN, VT and MLP modules for classification.•The performance of our CVM-Cervix exceeds that of existing methods and achieves a good result of 91.72% accuracy.

论文关键词：Convolutional neural network,Visual transformer,Multilayer perceptron,Cervical cell classification,Pap smear,Image classification

论文评审过程：Received 22 January 2022, Revised 15 May 2022, Accepted 2 June 2022, Available online 3 June 2022, Version of Record 7 June 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2022.108829