UVH-26: Urban Vision Hackathon Dataset for Indian Traffic Scenarios

UVH-26 is a large-scale, India-specific traffic-camera image dataset released by AIM @ IISc for research in intelligent transportation systems and vehicle-related computer vision. The dataset was last updated on December 10, 2025. Models trained on this data reportedly deliver up to 31.5% higher mAP than COCO-pretrained baselines.

Use Cases

• Train object detection models for vehicles and pedestrians based on traffic-camera imagery.
• Benchmark model performance in Indian traffic scenarios based on the dataset's real-world focus.
• Research intelligent transportation systems based on the dataset's stated purpose.
• Fine-tune pre-trained models for domain-specific gains based on the reported 31.5% mAP improvement.

Strengths

• Dataset is India-specific, providing geographically focused data.
• Models trained on this data demonstrate up to 31.5% higher mAP than COCO-pretrained baselines, suggesting strong domain relevance.
• Released by AIM @ IISc, a reputable research institution.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and license are unknown, which may limit suitability assessment.
• Data may reflect geographic bias inherent to its India-specific scope.

Provenance

Source

AIM @ IISc (Indian Institute of Science).

Collection Method

Likely collected from traffic-camera imagery.

Time Range

null

Freshness

Last updated 2025-12-10 09:33:48; freshness should be verified.

Geography

India.