Improving image captioning with Pyramid Attention and SC-GAN

作者：

Highlights：

• We proposed a variant of self-attention to extract the global features of the image.

• The visual and semantic relationship of different objects are integrated into our local-relation attention.

• We use a new self-critical training method to apply GAN to image captioning.

摘要

Highlights•We proposed a variant of self-attention to extract the global features of the image.•The visual and semantic relationship of different objects are integrated into our local-relation attention.•We use a new self-critical training method to apply GAN to image captioning.

论文关键词：Image captioning,Pyramid Attention network,Self-critical training,Reinforcement learning,Generative adversarial network,Sequence-level learning

论文评审过程：Received 18 September 2021, Revised 17 November 2021, Accepted 18 November 2021, Available online 30 November 2021, Version of Record 10 December 2021.

论文官网地址：https://doi.org/10.1016/j.imavis.2021.104340