Improving Baby-Food Safety with Hyperspectral Imaging & AI

How ELROILAB and Specim, Spectral Imaging Ltd teamed up to detect hidden contaminants and reduce quality issues by 90 %

ELROILAB, a South Korean deep-tech company, partnered with Specim to solve a critical challenge in baby-food production: detecting tiny bones, plastics and other foreign materials that traditional cameras and manual inspection often miss. Using Specim’s FX10 and FX17 hyperspectral cameras combined with ELROILAB’s SPECTRAL AI deep-learning platform, the system analyzes the spectral signature of every pixel – not just its appearance – enabling accurate identification of contaminants that look similar to the food material.

Integrated directly into the production line, the solution performs real-time inspection and automatically removes contaminated portions using air-blow, drop, or robotic vacuum systems. This hyperspectral, AI-powered approach reduced customer complaints by nearly 90 %, improved safety compliance and significantly lowered labour costs.

Chosen for its industrial reliability and superior spectral accuracy, Specim technology enables ELROILAB to deliver a highly scalable quality inspection solution that is now being expanded into broader food, recycling and manufacturing applications.