Image inpainting network for filling large missing regions using residual gather

作者：

Highlights：

• A multiple-level residual feature representation module is proposed.

• A novel image translation architecture is proposed.

• A hybrid loss including structure information is proposed to guide model training.

摘要

•A multiple-level residual feature representation module is proposed.•A novel image translation architecture is proposed.•A hybrid loss including structure information is proposed to guide model training.

论文关键词：Hole filling,Image inpainting,Residual learning,Generation model

论文评审过程：Received 10 December 2020, Revised 20 April 2021, Accepted 7 June 2021, Available online 15 June 2021, Version of Record 19 June 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115381