A face recognition dataset containing over 89,000 selfies and videos from more than 5,600 unique individuals. It was created by AxonData and was last updated on May 4, 2026. The dataset includes multi-ethnic subjects, varying lighting conditions, pose variations, and video recordings per person.
Use Cases
- Training facial recognition models based on the large collection of selfies and videos.
- Evaluating model performance on multi-ethnic subjects as described in the dataset.
- Testing model robustness under varying lighting conditions mentioned in the description.
- Developing algorithms for pose-invariant recognition using the pose variations included.
- Creating facial recognition databases using the video recordings per individual.
Strengths
- Contains over 89,000 selfies and video recordings.
- Includes more than 5,600 unique individuals.
- Features multi-ethnic subjects, varying lighting, and pose variations.
Limitations
- Description metadata is limited; actual data quality requires manual inspection after download.
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
Provenance
- Source
- AxonData
- Freshness
- Last updated 2026-05-04 16:45:58; freshness should be verified.