# dxli94/WLASL

WACV 2020 "Word-level Deep Sign Language Recognition from Video: A New Large-scale Dataset and Methods Comparison"

Repository: https://github.com/dxli94/WLASL
Canonical: https://ross.abutalabs.com/products/wlasl
Homepage: https://dxli94.github.io/WLASL/
Language: Python
License Family: other
Topics: sign-language-recognition, sign-language, sign-language-translation, sign-language-datasets, sign-language-classifier
Last push: 2023-03-18T15:55:21+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2424, "days_push": 1264, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1257, forks 159 (observed 2026-08-28T04:04:09.345272+00:00)

## What it is
WLASL is the largest video dataset for word-level American Sign Language (ASL) recognition, containing samples for 2,000 common ASL words/glosses with bounding box annotations. The repository provides dataset JSON files, download and preprocessing scripts, data loading code, and pre-trained I3D and Pose-TGCN models.

## Use cases
- train a sign language recognition model from video
- benchmark video classification models on ASL word recognition
- download and preprocess ASL video samples from YouTube
- research sign language understanding and translation
- fine-tune pretrained I3D or Pose-TGCN models on sign language data
- build an accessibility tool for deaf and hearing communication

## When to choose
- you need a large-scale word-level ASL video dataset for research
- you want benchmark baselines and pretrained models for sign language recognition
- you are working on video-based human action or gesture recognition
- you need gloss-level annotations with bounding boxes

## When to avoid
- you need continuous sign language sentence or full translation datasets rather than isolated words
- you need a commercially licensed dataset (C-UDA restricts to academic/computational use)
- you want non-American sign languages
- you cannot handle YouTube-dependent downloads or request missing videos from the authors

## Facets
- artifact type: dataset
- maturity: maintenance
- function: machine-learning, computer-vision, video-processing, data-science
- domain: computer-vision, machine-learning, accessibility
- platform: python, cross-platform
- tags: sign-language-recognition, asl, video-dataset, word-level-recognition, benchmark-dataset, c-uda-license, video

## Member repositories
- dxli94/WLASL (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:09.345272+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T05:07:10.971688+00:00, confidence not recorded.
  - readme: https://github.com/dxli94/WLASL (fetched 2026-08-28T04:04:09.345272+00:00, sha fa68d79ebe7a)
  - homepage: https://dxli94.github.io/WLASL/ (fetched 2026-08-29T12:17:40.997558+00:00, sha f4984dffd056)
- Data as of 2026-08-30T08:39:29.467469+00:00.