ArcFace Integration for React Native
Budget: $250 – $750 USD
I’ve built a React Native mobile app and now need to weave ArcFace / RetinaFace into it so each user can be reliably identified and presented with personalized content the moment they open the app. The goal isn’t gate-keeping or heavy security; it’s friction-free recognition that lets me tailor screens, offers, and settings to the person holding the phone.
You’ll take the pre-trained ArcFace or RetinaFace model, convert or wrap it for efficient on-device inference, and expose a clean React Native interface I can call from JavaScript. Accuracy and speed are critical, but the solution also has to respect mobile constraints—small footprint, low-latency camera access, and no noticeable battery drain.
Deliverables
• A plug-and-play React Native module (Android & iOS) that detects the face in real time, runs ArcFace / RetinaFace embedding, and returns a unique user ID.
• Simple enrollment and lookup functions so I can store vectors securely and match returning users.
• Sample screen demonstrating personalized content delivery after identification.
• Clear setup notes covering dependencies, model files, and any native build steps.
Acceptance criteria
• Identification latency under one second on mid-range devices.
• False-accept and false-reject rates in line with stock ArcFace benchmarks.
• Works offline; no server round-trip required.
• Passes a quick code review for readability and maintainability.
If you’ve already tinkered with ArcFace or RetinaFace on mobile—especially inside React Native—let’s talk. I’m happy to adjust scope later for extras like login authentication or user analytics, but for now the focus is rock-solid personalized content delivery.
You’ll take the pre-trained ArcFace or RetinaFace model, convert or wrap it for efficient on-device inference, and expose a clean React Native interface I can call from JavaScript. Accuracy and speed are critical, but the solution also has to respect mobile constraints—small footprint, low-latency camera access, and no noticeable battery drain.
Deliverables
• A plug-and-play React Native module (Android & iOS) that detects the face in real time, runs ArcFace / RetinaFace embedding, and returns a unique user ID.
• Simple enrollment and lookup functions so I can store vectors securely and match returning users.
• Sample screen demonstrating personalized content delivery after identification.
• Clear setup notes covering dependencies, model files, and any native build steps.
Acceptance criteria
• Identification latency under one second on mid-range devices.
• False-accept and false-reject rates in line with stock ArcFace benchmarks.
• Works offline; no server round-trip required.
• Passes a quick code review for readability and maintainability.
If you’ve already tinkered with ArcFace or RetinaFace on mobile—especially inside React Native—let’s talk. I’m happy to adjust scope later for extras like login authentication or user analytics, but for now the focus is rock-solid personalized content delivery.