.jpg)
AI vision can’t get better without learning from better images.
That’s why Perle Labs is powering a new data collection initiative designed to improve how AI systems recognize and understand human faces in the real world. This is part of a broader development effort supported by an enterprise partner in the field of computer vision, providing them with diverse and compliant data at scale through a decentralized workflow.
Contributors will be generating essential data for model training and evaluation. The goal is to build a verified dataset of real, diverse selfies that helps computer-vision systems recognize faces more accurately and fairly.
How to Contribute
Participation is open to whitelisted users who have joined the Perle Labs beta. Every project you complete adds to your rewards and reputation, unlocking higher-earning opportunities like this one. If you haven’t yet joined:
- Get onboarded.
- Complete available tasks.
- Become eligible.
If you’ve already onboarded, look out for new project notifications on your contributor dashboard. Contributors will be invited to participate in controlled batches on a rolling basis. Each batch will have a limited participation window to ensure equal opportunity.
The first batch launches this week.
What Participants Will Do
Contributors will submit a set of selfies that meet clear image guidelines for lighting, visibility, and framing. Before uploading, each participant completes a brief tutorial that walks through what qualifies as an acceptable photo (e.g., no filters, no masks, full face in view, etc.).
Each participant will then submit:
- A small number of new, on-the-spot selfies.
- A set of older selfies taken over time.
This combination of images will help models account for variables in lighting and appearance in order to strengthen the reliability of computer vision systems and ensure they work accurately.
Please Note: Quality matters. Data that doesn’t meet the requirements won’t earn rewards, and low-quality submissions can negatively impact your contributor reputation on the platform.
Real Impact, Real Rewards
Every quality submission supports active AI development aimed at making computer-vision systems more accurate and representative of human experience. This means that the data behind next-generation models reflects the diversity of real people.
This initiative also demonstrates how Perle Labs connects contributors with the teams and frontier labs pushing AI forward. Each completed task strengthens both the dataset and the network behind it, proving that better data starts with people earning fair rewards.
Follow @PerleLabs on X and join our Discord and Telegram communities to stay updated on batch releases and platform announcements.
