Marketplace
Computer VisionScore 91
Autonomous Driving Clips
Contributor0x3a4b...5c6d
ListedMay 8, 2026
Purchases12
About this Dataset
10,000 short video clips (3–8 seconds each) captured from front-facing and side cameras mounted on a vehicle driven through 6 major cities. Each frame is annotated with bounding boxes for pedestrians, vehicles, cyclists, traffic signs, and lane markings. Annotations were produced using a semi-automated pipeline with human verification at 10fps keyframe intervals. LiDAR point-cloud data is provided for 40% of clips. HD map integration metadata is included for all sequences.
Validation Report
Quality Analysis
Quality Score93 / 100
Issues
- !Night-scene coverage limited to 8% of total clips
Strengths
- Frame-accurate bounding box annotations at 10fps
- Multi-sensor data fusion (camera + LiDAR)
- High urban environment diversity across 6 cities
- HD map integration for ego-trajectory context
Category & Use Cases
Computer VisionVideo Object Detection for Autonomous Systems
Recommended Use Cases
Autonomous vehicle perception stacksPedestrian safety detectionTraffic sign and light recognitionLane-change prediction models
Originality Check
Similarity to Known Datasets9%
Proprietary capture routes with novel urban coverage. No meaningful intersection with NuScenes, KITTI, or Waymo Open Dataset.
Access Price
15.0OG
12
Total Purchases
May 8, 2026
Listed