Gaussian correction for adversarial learning of boundaries

作者：

Highlights：

• We aim to accurately segment faces from product videos for emotion classification.

• We propose a segmentation model that can be used on scenic images with occlusions.

• We consider a feedback loop that iteratively improves the results of segmentation.

摘要

•We aim to accurately segment faces from product videos for emotion classification.•We propose a segmentation model that can be used on scenic images with occlusions.•We consider a feedback loop that iteratively improves the results of segmentation.

论文关键词：Face expressions,Image segmentation,Discriminator loss,Gaussian correction

论文评审过程：Received 2 January 2022, Revised 13 June 2022, Accepted 9 August 2022, Available online 19 August 2022, Version of Record 1 September 2022.

论文官网地址：https://doi.org/10.1016/j.image.2022.116841