Textual-Visual Reference-Aware Attention Network for Visual Dialog | |
Guo, Dan1; Wang, Hui1; Wang, Shuhui2; Wang, Meng1 | |
刊名 | IEEE TRANSACTIONS ON IMAGE PROCESSING |
2020 | |
卷号 | 29页码:6655-6666 |
关键词 | Visual dialog attention network textual reference visual reference multimodal semantic interaction |
ISSN号 | 1057-7149 |
DOI | 10.1109/TIP.2020.2992888 |
英文摘要 | Visual dialog is a challenging task in multimedia understanding, which requires the dialog agent to answer a series of questions that are based on an input image. The critical issue to produce an exact answer is how to model the mutual semantic interaction among feature representations of the image, question-answer history, and current question. In this study, we propose a textual-visual Reference-Aware Attention Network (RAA-Net), which aims to effectively fuse Q (question), H (history), V-l (local vision), and V-g (global vision) to infer the exact answer. In the multimodal feature flows, RAA-Net first learns the textual context through multi-head attention between Q and H and then guides the textual reference semantics to the image to capture visual reference semantics by self- and cross-reference-aware attention in and between V-l and V-g. In the proposed RAA-Net, we exploit the two-stage (intra- and inter-) visual reasoning mechanism on V-l and V-g. Extensive experiments on the VisDial v0.9 and v1.0 datasets show that RAA-Net achieves state-of-the-art performance. Visualization results on both visual and textual attention maps further validate the remarkable interpretability achieved by our solution. |
资助项目 | National Natural Science Foundation of China (NSFC)[61725203] ; National Natural Science Foundation of China (NSFC)[61732008] ; National Natural Science Foundation of China (NSFC)[61672497] ; National Natural Science Foundation of China (NSFC)[61876058] |
WOS研究方向 | Computer Science ; Engineering |
语种 | 英语 |
出版者 | IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC |
WOS记录号 | WOS:000545739000006 |
内容类型 | 期刊论文 |
源URL | [http://119.78.100.204/handle/2XEOYT63/15073] |
专题 | 中国科学院计算技术研究所期刊论文_英文 |
通讯作者 | Wang, Meng |
作者单位 | 1.Hefei Univ Technol, Sch Comp Sci & Informat Engn, Hefei 230601, Peoples R China 2.Chinese Acad Sci, Inst Comp Technol, Key Lab Intelligent Informat Proc, Beijing 100190, Peoples R China |
推荐引用方式 GB/T 7714 | Guo, Dan,Wang, Hui,Wang, Shuhui,et al. Textual-Visual Reference-Aware Attention Network for Visual Dialog[J]. IEEE TRANSACTIONS ON IMAGE PROCESSING,2020,29:6655-6666. |
APA | Guo, Dan,Wang, Hui,Wang, Shuhui,&Wang, Meng.(2020).Textual-Visual Reference-Aware Attention Network for Visual Dialog.IEEE TRANSACTIONS ON IMAGE PROCESSING,29,6655-6666. |
MLA | Guo, Dan,et al."Textual-Visual Reference-Aware Attention Network for Visual Dialog".IEEE TRANSACTIONS ON IMAGE PROCESSING 29(2020):6655-6666. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论