0

ConTra: (Con)text (Tra)nsformer for Cross-Modal Video Retrieval

Context Transformer (ConTra) enhances cross-modal clip-sentence retrieval by modeling local temporal context with contrastive losses, improving performance on video datasets.

Year
2022
Venue
arXiv 2022
Authors
3
Hosting
Abstract onlyARXIV-DEFAULT

Cite

Notes

Only stored in your browser.

Attribution

Abstract & full text
arxiv.org/abs/2210.04341ARXIV-DEFAULT
TL;DR
Semantic Scholar
Attribution policy →

Abstract

In this paper, we re-examine the task of cross-modal clip-sentence retrieval, where the clip is part of a longer untrimmed video. When the clip is short or visually ambiguous, knowledge of its local temporal context (i.e. surrounding video segments) can be used to improve the retrieval performance. We propose Context Transformer (ConTra); an encoder architecture that models the interaction between a video clip and its local temporal context in order to enhance its embedded representations. Importantly, we supervise the context transformer using contrastive losses in the cross-modal embedding space. We explore context transformers for video and text modalities. Results consistently demonstrate improved performance on three datasets: YouCook2, EPIC-KITCHENS and a clip-sentence version of ActivityNet Captions. Exhaustive ablation studies and context analysis show the efficacy of the proposed method.

Authors

3