Challenge
Reflective gold, gemstones, and variable lighting make automated authentication difficult. Gold Guard needed evidence that a handheld authentication device was technically achievable, plus a credible investor roadmap.
Solution
We combined macro imaging with cross-polarised and coaxial lighting, IMU-gated capture for stable images, and a quantised edge-AI verification model, defining laboratory testing, hardware direction, and a delivery roadmap.
Results
- Specialised lighting reduced reflections and revealed key details
- Motion sensors ensured stable image capture
- A lightweight on-device AI model returns a result in under five seconds
- Defined lab tests, hardware requirements, development stages, and investment roadmap
Technology
Edge AI · Computer Vision · Quantised Inference
Powered by DAC Digital