Weather-degraded image semantic segmentation with multi-task knowledge distillation

作者：

Highlights：

• We propose a multi-task knowledge distillation method.

• Create a system that combines feature distillation and structure distillation.

• We propose a shift operator to realize lightweight model.

• We build a new dataset called Snowy-Cityscapes.

摘要

•We propose a multi-task knowledge distillation method.•Create a system that combines feature distillation and structure distillation.•We propose a shift operator to realize lightweight model.•We build a new dataset called Snowy-Cityscapes.

论文关键词：Adverse weather,Road scene,Knowledge distillation,Image enhancement,Semantic segmentation

论文评审过程：Received 7 June 2022, Revised 27 August 2022, Accepted 9 September 2022, Available online 15 September 2022, Version of Record 21 September 2022.

论文官网地址：https://doi.org/10.1016/j.imavis.2022.104554