eduzhai > Applied Sciences > Engineering >

Graph Neural Network for Video-Query based Video Moment Retrieval

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 29 pages

Abstract: In this paper, we focus on Video Query based Video Moment Retrieval (VQ-VMR)task, which uses a query video clip as input to retrieve a semantic relativevideo clip in another untrimmed long video. we find that in VQ-VMR datasets,there exists a phenomenon showing that there does not exist consistentrelationship between feature similarity by frame and feature similarity byvideo, which affects the feature fusion among frames. However, existing VQ-VMRmethods do not fully consider it. Taking this phenomenon into account, in thisarticle, we treat video features as a graph by concatenating the query videofeature and proposal video feature along time dimension, where each timestep istreated as a node, each row of the feature matrix is treated as feature of eachnode. Then, with the power of graph neural networks, we propose a Multi-GraphFeature Fusion Module to fuse the relation feature of this graph. Afterevaluating our method on ActivityNet v1.2 dataset and Thumos14 dataset, we findthat our proposed method outperforms the state of art methods.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...