Hand pose and action recognition data from real Nigerian work.
First-person video of skilled hands at work, with 21-keypoint hand pose and action segments, filmed in real workshops, salons, kitchens and markets. It adds skin tones, tools and settings most hand datasets lack.
Free 10-hour sample first. A person replies within one working day.
Footage that suits this use case.
Where the data runs short.
Skin tone and lighting
Hand and pose models can perform worse on darker skin and in uneven light. Training data needs to include them.
Tools hide fingers
Hands holding tools are partly occluded. Models need many labelled examples of that, not just open hands.
Fine-grained verbs
Action recognition needs exact start and end times for actions such as “cut”, “thread”, “tighten”, not one label per video.
What you get from TrustAm Data.
- Hand pose
- 21 keypoints per hand per labelled frame, with visibility flags
- Action segments
- Start and end time, verb, object and tool per segment
- Captions
- Short free-text description per segment, in English
- Review
- Automatic checks, then a person reviews every minute and samples labels
- Diversity
- Dark-skinned hands, indoor and outdoor light, many tools and materials
Scope it so the first sample is useful.
- Name the action vocabulary you need, or start from ours and edit it
- Choose a labelled-frame rate for hand pose (for example every frame, or a sampled rate)
- Tell us the trades with the hand actions you care about
- Agree a label QA target in the free sample before the pilot
Provenance on every minute.
ID-checked recorders
BVN or NIN plus a live selfie matched to the ID, and a bank account in the same name, before any paid task.
Consenting workplaces
Each workplace is registered in person by a TrustAm field agent, with the owner’s consent and a location fence.
Live capture only
A spoken challenge code, motion sensors and GPS show each clip was filmed live, at work. No camera-roll uploads.
Blurred and reviewed
Bystander faces and number plates are blurred; a trained reviewer approves or cuts every minute before delivery.
Buyers usually ask.
Who labels the hand pose?
Labels start with a model pass and are corrected and checked by trained reviewers. We agree a quality target with you during the free sample.
Can you deliver unlabelled video only?
Yes. Curated minutes without labels are the simplest layer; you can add hand pose or action segments later on the same footage.
How do I start?
Email [email protected] or use the request form. A person replies within one working day and we record a free sample of about 10 hours to your spec. Pricing is quoted per approved hour after scoping.
More from TrustAm Data.
Tell us what you need. We send a free sample first.
Email [email protected] with the data type, hours, languages or trades, labels and format. Pricing is quoted per approved hour after scoping.