ArCo: Attention-reinforced transformer with contrastive learning for image captioning
Image and Vision Computing(2022)
摘要
•This paper present an attention-based method that uses contrastive learning.•The loss function is a combination of CE loss and contrastive learning loss.•This paper shown to exhibit state-of-the-art performances on well-known datasets.
更多查看译文
关键词
Image captioning,Visual attention,Transformer,Contrastive learning
AI 理解论文
溯源树
样例
![](https://originalfileserver.aminer.cn/sys/aminer/pubs/mrt_preview.jpeg)
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要