Nov 3, 2022 · Our three-step approach consists of moment proposal, moment-query matching and postprocessing, all using only off-the-shelf models. On the ...

We propose a zero-shot simple approach for one such task, Video Moment Retrieval (VMR), that does not perform any additional finetuning and simply repurposes ...

This work proposes a zero-shot simple approach for Video Moment Retrieval that does not perform any additional finetuning and simply repurposes ...

Our three-step approach consists of moment proposal, moment-query matching and postprocessing, all using only off-the-shelf models. On the QVHighlights ...

Accurate video moment retrieval (VMR) requires uni- versal visual-textual correlations that can handle unknown vocabulary and unseen scenes.

Video Moment Retrieval, which aims to locate in-context video moments according to a natural language query, is an essential task for cross-modal grounding.

Dec 1, 2023 · We show that such models can be easily repurposed as effective, off-the-shelf feature extractors for VMR. On the QVHighlights benchmark for VMR, ...

... Zero-shot video moment retrieval with off-the-shelf models. CoRR abs/2211.02178 (2022). https://doi.org/10.48550/arXiv.2211.02178 https://doi.org/10.48550 ...

Video moment localization, also known as video moment retrieval, aiming ... Zero-shot Video Moment Retrieval With Off-the-Shelf Models. no code yet • 3 ...

Given a video and a natural language query, the task of Video Moment Retrieval (VMR) involves temporally localizing moments (video segments) within the ...