How Overview AI and Advantech Bring AI Vision Inspection to the Factory Floor

A look at Overview AI deep learning models running on Advantech ICAM, plus where to see the live setup at Computex 2026 in Taipei.
Visual inspection in manufacturing is finally catching up to AI. After years of cameras living in one stack and AI models living in another, the integrations that actually deliver edge AI vision inspection are starting to ship. At Computex 2026, Overview AI and Advantech are bringing one of those integrations to the floor.
The setup running at the Advantech booth at TaiNEX Hall 1 brings two industrial-grade components together as one production pipeline.
Advantech ICAM
An industrial camera platform built for production environments, with consistent capture, the form factor to mount in real fixtures, and the connectivity to deliver clean frames to downstream compute.
Overview AI
The Overview AI inspection model runs against each frame, classifying parts and surfacing quality conditions within the camera cycle time. Trained for the specific products on a line, updated by a quality engineer without rewriting integrations. Structured pass and fail events feed the factory MES, SCADA, or quality data layer.
The combination is the kind of vision inspection pipeline quality engineers have been asking for. No cloud round-trip. No quarter-long integration project. No specialized vision team required to deploy it.

The Overview AI OV20i, the same camera class used in the live demo at Computex 2026.
Where to see the integration live
The Overview AI and Advantech setup runs at the Advantech booth at TaiNEX Hall 1 throughout Computex 2026. The Advantech booth Stage Show featuring Overview AI takes place on June 2, starting at 1:00 PM, with the live demo on real parts. A second demo runs at the Advantech Linkou Campus during the WPC IIoT Sector Forum on June 3.
Two sessions from the Overview AI team
Xiao Kao, Co-founder of Overview AI
Speaking at the WPC IIoT Sector Forum on production deployments of deep learning inspection across electronics, automotive, and AI infrastructure manufacturing, with specific patterns from lines running thousands of inspections per shift.
Russell Nibbelink, Co-founder and COO of Overview AI
Sharing how the latest generation of AI tooling, including synthetic image generation, is cutting visual inspection deployment time from weeks to a single day. The talk is grounded in deployments where new product introduction used to push inspection bring-up into multi-week territory, and where Gen AI shifted that work into a single quality engineer shift.
Why this matters for Taiwan manufacturers
Taiwan electronics manufacturers run some of the most demanding visual inspection workloads in the world, with high mix, high volume, and a zero-escape quality bar. The Overview AI and Advantech integration is built for exactly that environment. The ICAM holds up to the floor. The Overview AI model holds up to the inspection bar. The pipeline holds up to a shift.
If your team is running inspection at scale, stop by the Advantech booth at TaiNEX Hall 1 or the WPC IIoT Sector Forum at Linkou. We will walk you through the live demo and the deployment path for your own lines.
See the live setup
Advantech booth · TaiNEX Hall 1 · Computex 2026
June 2 at 1:00 PM for the stage show. Or the WPC IIoT Sector Forum at the Advantech Linkou Campus on June 3.
Talk to the Overview AI teamAbout the companies
About Overview AI
Overview AI is an industrial AI vision platform deployed across more than 2,000 cameras in factories on four continents, covering electronics, automotive, AI infrastructure, aerospace, medical devices, and batteries. Overview cameras (OV20i, OV80i) combine an industrial fanless form factor with on-device deep learning inference.
About Advantech
Advantech is a global leader in industrial IoT. The ICAM industrial camera platform is part of Advantech’s broader edge intelligence portfolio used across factory automation deployments worldwide.
Related Reading
More on Overview AI deployments and edge AI vision inspection: