AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

BSL-1K: Scaling up co-articulated sign language recognition using mouthing cues

Samuel Albanie, Gül Varol, Liliane Momeni, Triantafyllos Afouras, Joon Son Chung, Neil Fox, Andrew Zisserman

ECCV 2020 · Large language models › Language, safety and interpretability · approach: Sign language datasets and translation · 195 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
lower mouthing threshold to 0.5 (more, noisier training data) (ablation)I3D Kinetics init · BSL-1K test (manually verified)per-instance top-1 accuracy: 69.00 -> 70.61Improved
add body and hand cues to face (full-frame instead of face-crop) (ablation)I3D · BSL-1K m.8 pose subsetper-instance top-1 accuracy: 42.23 -> 65.57Improved
add the face to body-and-hand keypoints (ablation)Pose->Sign · BSL-1K m.8 pose subsetper-instance top-1 accuracy: 40.47 -> 49.66Improved
use I3D on raw video instead of 2D pose (ablation)I3D vs Pose->Sign · BSL-1K m.8 pose subsetper-instance top-1 accuracy: 49.66 -> 65.57Improved
pretrain with video pose distillation instead of random init (ablation)I3D · BSL-1K m.8per-instance top-1 accuracy: 39.80 -> 70.38Improved
pretrain with video pose distillation instead of Kinetics (ablation)I3D · BSL-1K m.8per-instance top-1 accuracy: 69.00 -> 70.38Improved
pretrain on gesture recognition (Jester) instead of random init (ablation)I3D · BSL-1K m.8per-instance top-1 accuracy: 39.80 -> 46.93Improved
use 8 sec keyword search window instead of 1 sec (ablation)I3D Kinetics init · BSL-1K m.8per-instance top-1 accuracy: 60.10 -> 69.00Improved
use 8 sec keyword search window instead of 16 sec (ablation)I3D Kinetics init · BSL-1K m.8per-instance top-1 accuracy: 65.91 -> 69.00Improved
use 20 frames before mouthing peak instead of 16 (ablation)I3D Kinetics init · BSL-1K m.8per-instance top-1 accuracy: 59.53 -> 71.71Improved
use 20 frames before mouthing peak instead of 24 (ablation)I3D Kinetics init · BSL-1K m.8per-instance top-1 accuracy: 69.00 -> 71.71Improved
mask mouth at test time for a model trained on full frames (ablation)I3D · BSL-1K m.8 pose subsetper-instance top-1 accuracy: 65.57 -> 34.74Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy