WLASL (World Level American Sign Language) Video

• WLASL_v0.3.json: Contains the glossary and video instance details required for processing.
• Videos Folder: Comprises approximately 12k videos, each named according to its unique video_id.

Usage:

This dataset is ideal for:

• Word-Level Action Recognition: Classifying ASL words into textual representations.
• Sign Language Research: Facilitating the development of models for sign language understanding.
• Computer Vision Applications: Training and testing models for word-level sign language recognition.

Acknowledgments:

The WLASL dataset is created by Dongxu Li and Hongdong Li and is licensed under the Computational Use of Data Agreement (C-UDA). Please refer to the C-UDA-1.0 license for detailed terms.

Inspiration:

• How can we classify word-level ASL actions into text accurately?
• What are the best-performing models for word-level sign language recognition?

License and Updates:

• License: Computational Use of Data Agreement (C-UDA)
• Expected Update Frequency: Never