NIA-Network: Towards improving lung CT infection detection for COVID-19 diagnosis

作者：

Highlights：

• A new semi-supervised method is proposed to improve COVID-19 detection performance.

• This method relies on Network-in-Network and Instance Normalization to normalize CT.

• We utilize adversarial learning to transfer labels from source to target domain.

• We utilize regional proposal network to learn the domain-invariant detected regions.

摘要

•A new semi-supervised method is proposed to improve COVID-19 detection performance.•This method relies on Network-in-Network and Instance Normalization to normalize CT.•We utilize adversarial learning to transfer labels from source to target domain.•We utilize regional proposal network to learn the domain-invariant detected regions.

论文关键词：COVID-19 diagnosis,Semi-supervised learning,Adversarial learning,Network-in-Network,Instance normalization

论文评审过程：Received 20 June 2020, Revised 12 April 2021, Accepted 26 April 2021, Available online 2 May 2021, Version of Record 26 May 2021.

论文官网地址：https://doi.org/10.1016/j.artmed.2021.102082