Visual defect inspection — Pharmaceutical
Deep-learning vision that flags surface and assembly defects at line speed. Delivered for a pharmaceutical client in Kutch.
- Location
- Kutch
- Year
- 2022
- Duration
- 18 weeks
- Team
- 9 engineers
- 98%
- defect recall
- 28
- ms per part
- 92%
- fewer escapes
The challenge
The pharmaceutical team was relying on manual effort that could not keep up with volume, and every miss had a direct cost.
What we built
We ran a two-week discovery sprint, then built and fielded the system end to end on PyTorch, TensorRT and OpenCV, with operator training and an SLA.
Delivery timeline
Phase 1
Discovery sprint and site survey
Phase 2
Prototype built and tested in the Motera lab
Phase 3
Pilot on site with before/after measurement
Phase 4
Scale-out, training and handover
Highlights
- Documentation, drawings and source handed over at close
- Designed, built and supported from Motera, Ahmedabad
- Before-and-after measurements agreed with the client up front
- Safety case reviewed with the client's EHS team
Related AI projects
AIDeliveredAutomotive · Mundra · 2020
Visual defect inspection — Automotive
Deep-learning vision that flags surface and assembly defects at line speed. Delivered for a automotive client in Mundra.
- 98%
- defect recall
- 27
- ms per part
- 84%
- fewer escapes
AILiveTextile · Bharuch · 2024
Visual defect inspection — Textile
Deep-learning vision that flags surface and assembly defects at line speed. Delivered for a textile client in Bharuch.
- 97%
- defect recall
- 15
- ms per part
- 61%
- fewer escapes
AIDeliveredElectronics · Gandhinagar · 2020
Visual defect inspection — Electronics
Deep-learning vision that flags surface and assembly defects at line speed. Delivered for a electronics client in Gandhinagar.
- 100%
- defect recall
- 30
- ms per part
- 71%
- fewer escapes
AIPilotCeramics · Anand · 2021
Visual defect inspection — Ceramics
Deep-learning vision that flags surface and assembly defects at line speed. Delivered for a ceramics client in Anand.
- 96%
- defect recall
- 40
- ms per part
- 61%
- fewer escapes
Have a problem that needs a machine built for it?
Tell us the constraint — line speed, accuracy, environment, deadline. An engineer will read it and reply with a straight answer, including when the answer is that you don't need us.
We reply to every enquiry within one business day.