Training data for AI that works in the real world
Our teams are trained per domain, so the people labelling your data already know what a lane marking, a lesion or a pallet should look like.
Autonomous vehicles & ADAS
Road-scene annotation for perception models: vehicles, pedestrians, cyclists, lanes and traffic signs.
- 2D boxes and 3D cuboids on camera images
- Lane and drivable-area segmentation
- Multi-frame object tracking
- Traffic sign and light classification
Lane-mark and road-damage annotation on 5,000 road images, plus weather and lighting classification on 10,000+, for a US autonomous driving company.
Robotics
Training data for robots that see, grasp and move in warehouses, factories and outdoors.
- Multimodal annotation for fleet navigation
- Object detection and segmentation for picking
- Scene segmentation for navigation
- Keypoints and pose
- Model output review for edge cases
Navigable-space polygons on 20,000 images and amodal annotation on 5,000 more for a fleet of quadruped and wheeled robots.
Healthcare & medical imaging
Segmentation for diagnostic AI across CT, MRI and X-ray.
- Organ segmentation
- Structure and lesion outlining
- Cross-modality guidelines
- Two-layer QA
Geospatial
Satellite, aerial and drone imagery labelled for mapping, monitoring and land-use models.
- Building and road segmentation
- Land-cover classification
- Object detection in aerial images
- Change detection
Location-type polygons, rooftops and solar panels on about 5,000 drone images for an Austrian drone company.
Agriculture & forestry
Tree, crop and species detection from field, drone and aerial images.
- Tree and species detection
- Crop and plant segmentation
- Field image classification
Sports & media
Player, ball and event tracking across match footage for analytics and automated production.
- Player and object tracking
- Event time-stamping
- Pose and keypoints
- Embedded long-term teams
Retail & e-commerce
Product, shelf and catalogue data for visual search, inventory and recommendation models.
- Product detection and tagging
- Shelf and planogram annotation
- Attribute labelling for catalogues
- Product text and review classification
Product and shelf image annotation for retail and e-commerce models.
Working in another field?
If your model learns from images, video, text or audio, we can likely help. Start with a free pilot batch.