Data augmentation for skin lesion using self-attention based progressive generative adversarial network

作者：

Highlights：

• Proposing self-attention progressive growing of generative adversarial networks.

• Applying an imbalanced learning rate for increased stability of training behavior.

• Synthesizing high-quality and fine-grained skin lesion images.

• Improving melanoma detection of automated skin lesion detection.

摘要

•Proposing self-attention progressive growing of generative adversarial networks.•Applying an imbalanced learning rate for increased stability of training behavior.•Synthesizing high-quality and fine-grained skin lesion images.•Improving melanoma detection of automated skin lesion detection.

论文关键词：Skin cancer,Generative models,Deep learning,Data augmentation,Data imbalance

论文评审过程：Received 4 April 2020, Revised 23 August 2020, Accepted 25 August 2020, Available online 3 September 2020, Version of Record 9 September 2020.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.113922