SLSNet: Skin lesion segmentation using a lightweight generative adversarial network
作者：
Highlights：
• A lightweight and fully automatic skin lesion segmentation model is proposed.
• A multiscale mechanism is introduced to extract features at different scales.
• The position attention module controls the spatial inter-dependencies.
• The channel attention module controls the channel inter-dependencies.
• The combination of binary cross-entropy, Jaccard index, and L1 loss is used.
摘要
•A lightweight and fully automatic skin lesion segmentation model is proposed.•A multiscale mechanism is introduced to extract features at different scales.•The position attention module controls the spatial inter-dependencies.•The channel attention module controls the channel inter-dependencies.•The combination of binary cross-entropy, Jaccard index, and L1 loss is used.
论文关键词：Skin lesion segmentation,Deep generative adversarial network,1-D kernel factorized network,Position attention,Channel attention
论文评审过程：Received 10 July 2020, Revised 14 May 2021, Accepted 10 June 2021, Available online 17 June 2021, Version of Record 19 June 2021.
论文官网地址：https://doi.org/10.1016/j.eswa.2021.115433
