Computer Vision Toolbox - Connected Computer Vision & AI Quality Inspection at Scale!
summary
ZF builds an in‑house Computer Vision Toolbox: an edge agent that triggers cameras and sends images + metadata to the cloud over DMP, stores and trains in AAP and saves models in a model registry for reuse. The platform targets 188 production sites (95 EU, 32 Asia, 31 NA), archives images in a corporate data lake, enforces camera/configuration management, deploys models via automated pipelines, monitors instances for observability, and applies MLOps (process, people, tools, observability, pipelines, documentation) to shorten deployment from years to weeks.
AI-written summary · about
chapters (11)
- 0:00Computer Vision Toolbox intro
- 1:09Global quality inspection challenges
- 4:02Limitations in automated inspection
- 6:29Toolbox architecture & components
- 9:40Configurability & deployment
- 11:05Standardized solutions for roadblocks
- 12:48MLOps for AI deployment
- 15:58Summary & future vision
- 16:45Q&A: Model reuse
- 18:33Q&A: Competing with suppliers
- 21:43Q&A: Data collection
related talks
27mComputer vision for automated quality inspection
36mAI-Powered Real-Time Quality Control
41mPractical Aspects of Applying Modern Computer Vision Technologies
33mBuilding end-to-end computer vision solutions from deep learning models
43mIntersection of Awesome: Building AI enabled IoT solutions powered by the cloud
01h 33mComputer Vision in Azure
29mApplied Custom Vision
14mComputer vision on edge devices