This video presentation describes the paper:
Samuel Albanie*, Gül Varol*, Liliane Momeni, Triantafyllos Afouras, Joon Son Chung, Neil Fox and Andrew Zisserman, "BSL-1K: Scaling up co-articulated sign language recognition using mouthing cues", ECCV 2020.
Summary:
1) We automatically annotate:
- 1,000 hours of continuous sign language videos,
- using visual keyword spotting,
- for a vocabulary of 1,064 words.
2) We train a sign recognition spatio-temporal network using the automatic labels;
3) We test the network on manually verified signs.
arxiv: https://arxiv.org/abs/2007.12131
