English

Visual Keyword Spotting with Attention

Computer Vision and Pattern Recognition 2021-11-01 v1 Computation and Language

Abstract

In this paper, we consider the task of spotting spoken keywords in silent video sequences -- also known as visual keyword spotting. To this end, we investigate Transformer-based models that ingest two streams, a visual encoding of the video and a phonetic encoding of the keyword, and output the temporal location of the keyword if present. Our contributions are as follows: (1) We propose a novel architecture, the Transpotter, that uses full cross-modal attention between the visual and phonetic streams; (2) We show through extensive evaluations that our model outperforms the prior state-of-the-art visual keyword spotting and lip reading methods on the challenging LRW, LRS2, LRS3 datasets by a large margin; (3) We demonstrate the ability of our model to spot words under the extreme conditions of isolated mouthings in sign language videos.

Keywords

Cite

@article{arxiv.2110.15957,
  title  = {Visual Keyword Spotting with Attention},
  author = {K R Prajwal and Liliane Momeni and Triantafyllos Afouras and Andrew Zisserman},
  journal= {arXiv preprint arXiv:2110.15957},
  year   = {2021}
}

Comments

Appears in: British Machine Vision Conference 2021 (BMVC 2021)

R2 v1 2026-06-24T07:18:19.176Z