ASL Citizen Processed Keypoints-200: Isolated American Sign Language Recognition Data

Processed keypoints from the ASL Citizen dataset, intended for isolated sign recognition and encoder pretraining. The dataset was created by SharoonArshad and last updated on May 17, 2026. It is designed as the first stage of a larger pipeline for translating continuous American Sign Language to English.

Use Cases

• Train an isolated sign encoder based on processed keypoint data.
• Pretrain an encoder for transfer into a Connectionist Temporal Classification model.
• Fine-tune a model on continuous signing data, such as How2Sign, after initial training on this dataset.
• Develop components for an ASL-to-English translation pipeline.

Strengths

• Dataset is specifically processed for isolated sign recognition and encoder pretraining, indicating a clear methodological purpose.
• The description outlines a concrete pipeline for use in continuous sign language translation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and license information are unknown, which may limit suitability assessment.
• Data may reflect bias inherent to the original source dataset.

Provenance

Source

Processed from the ASL Citizen keypoints-200 Kaggle dataset.

Collection Method

Processing method is not specified in the provided description.

Freshness

Last updated 2026-05-17 19:19:17; freshness should be verified.