arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1909.02701cs.CV
收藏

面向图文匹配的视觉语义推理

Visual Semantic Reasoning for Image-Text Matching

Northeastern University(东北大学)

Kunpeng Li, Yulun Zhang, Kai Li, Yuanyuan Li, Yun Fu

摘要:

图文匹配一直是连接视觉和语言领域的研究热点。它仍然具有挑战性,因为目前的图像表示通常缺乏全局语义概念,在其相应的文本标题。为了解决这个问题,我们提出了一个简单的和可解释的推理模型,以产生视觉表示,捕捉关键对象和语义概念的场景。具体来说,我们首先建立图像区域之间的连接,并执行推理与图卷积网络生成具有语义关系的功能。然后,我们建议使用门和记忆机制,这些关系增强功能进行全局语义推理,选择的歧视性信息,并逐步产生整个场景的表示。实验结果表明,该方法在MS-COCO和Flickr 30 K数据集上实现了一种新的图文匹配方法。在MS-COCO(Recall@1使用1 K测试集)上,该方法在图像检索和字幕检索方面分别比目前最好的方法提高了6.8%和4.8%。在Flickr 30 K上,我们的模型相对提高了12.6%的图像检索和5.8%的字幕检索(Recall@1)。我们的代码可在https://github.com/KunpengLi1994/VSRN上获得。

英文摘要:

Image-text matching has been a hot research topic bridging the vision and language areas. It remains challenging because the current representation of image usually lacks global semantic concepts as in its corresponding text caption. To address this issue, we propose a simple and interpretable reasoning model to generate visual representation that captures key objects and semantic concepts of a scene. Specifically, we first build up connections between image regions and perform reasoning with Graph Convolutional Networks to generate features with semantic relationships. Then, we propose to use the gate and memory mechanism to perform global semantic reasoning on these relationship-enhanced features, select the discriminative information and gradually generate the representation for the whole scene. Experiments validate that our method achieves a new state-of-the-art for the image-text matching on MS-COCO and Flickr30K datasets. It outperforms the current best method by 6.8% relatively for image retrieval and 4.8% relatively for caption retrieval on MS-COCO (Recall@1 using 1K test set). On Flickr30K, our model improves image retrieval by 12.6% relatively and caption retrieval by 5.8% relatively (Recall@1). Our code is available at https://github.com/KunpengLi1994/VSRN.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑