Computer vision for automated quality inspection
summary
Claims that Amazon Lookout for Vision builds usable custom defect models as few as 30 images, runs inference on-device like an Nvidia Jetson edge inference, and integrates with cloud services to feed a centralized OEE dashboard. Case results include Invista cutting inspection cycle time from 60s to 5s, Tyson replacing manual checks over 8,000 pins, and a demonstrator that uses a robotic arm to divert defective bottles in real time, proving the pipeline works end to end.
AI-written summary · about
chapters (10)
- 0:00Introduction to talk
- 1:04Quality management importance
- 3:35Computer vision in quality
- 8:18Amazon Lookout for Vision overview
- 12:30Model training and MLOps
- 14:17Common use cases
- 15:44Customer success stories
- 19:21Quality control at scale
- 22:33Case study demo
- 26:27Summary and Q&A
related talks
29mAmazon Lookout for Vision
14mComputer vision on edge devices
36mAI-Powered Real-Time Quality Control
17mAI in Quality Control – Real Use Cases and Best Practices
23mComputer Vision Toolbox - Connected Computer Vision & AI Quality Inspection at Scale!
41mPractical Aspects of Applying Modern Computer Vision Technologies
01h 00mOpening Keynote
26mServerless Architecture for Product Defect Detection