- Client
- Proditec
- Type
- Technology transfer project
- Pillar
- Consolidate
- Duration
- December 2021 to December 2023
- Official website
- Proditec
The challenge
Proditec designs high-speed sorting and measuring machines for the pharmaceutical and currency industries. The company wanted to build its own skills in AI-based computer vision, with a tough constraint: algorithms that run in real time, in production conditions.
CATIE’s role
- A hybrid vision station able to image capsules and tablets and reveal defects
- State-of-the-art AI algorithms to classify images by type of defect, with the best balance between speed and accuracy
- Selection of an embedded target offering the best speed-accuracy trade-off
- Deployment of the algorithm on that target
Expected benefits
- Overcome challenges in electronics and AI applied to industrial image processing
- Build on these advances to design future vision stations across the entire Proditec range
- Give the company in-house AI vision skills
Expertise involved
This project received support from the Nouvelle-Aquitaine Region.
