Visual qualityinspection
Camera-based inspection for med-tech production lines, pairing learned anomaly detection with rule-based checks you can validate.
Concept illustration – not real results
Inspection you can put in a validation report
Manual visual inspection is tiring, inconsistent between shifts and hard to document. Black-box AI inspection is fast but difficult to justify to an auditor.
We combine both worlds: deterministic measurements where a rule exists, and anomaly detection for the defects nobody wrote a rule for.
Anomaly detection
Learns what a good part looks like and flags deviations, without needing a large library of defect images.
Rule-based checks
Dimensions, presence and alignment measured deterministically, with tolerances you set and can trace.
Validation evidence
Every decision is logged with the image and the reason, ready for your process validation.
Rendered concept illustration – parts, measurements and labels are fictitious.
Specification
- Camera
- To be defined
- Throughput
- To be defined
- Detection
- Anomaly detection + rule-based checks
- Interface
- To be defined
- Deployment
- On-premise
- Documentation
- To be defined
Where it stands
- Demo softwareAnomaly detection and non-AI checks, in development
- Demo hardwareLow-cost rig, planned
- Trade-fair demonstratorPlanned
See it on your parts
Send us a few good and defective samples and we will show what the system sees and how it decides.