SAR-to-optical image translation based on improved CGAN

作者：

Highlights：

• We propose an improved CGAN method for SAR-to-optical image translation.

• We propose a parallel feature fusion generator to improve the contour sharpness.

• We utilize a multi-scale discriminator to improve the texture fine-grainedness.

• We design a chromatic aberration loss to improve the color fidelity of the generated optical image.

摘要

•We propose an improved CGAN method for SAR-to-optical image translation.•We propose a parallel feature fusion generator to improve the contour sharpness.•We utilize a multi-scale discriminator to improve the texture fine-grainedness.•We design a chromatic aberration loss to improve the color fidelity of the generated optical image.

论文关键词：SAR-to-optical image translation,Chromatic aberration loss,ICGAN

论文评审过程：Received 26 February 2021, Revised 16 June 2021, Accepted 27 July 2021, Available online 2 August 2021, Version of Record 8 August 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108208