Leica Microsystems and Cognex announced Aivaro AI Feature Detection on September 29. The module uses Cognex vision technology inside Leica's Enersight platform to identify, classify and measure features during microscope-based inspection; the companies say processing can remain on-site.
GLOBAL IMAGING BRIEF
Analysis: connecting image capture, settings and defect classification could make inspection decisions more repeatable across shifts. Training-image requirements and accuracy still need to be validated for each material, defect class and production environment.
Full perspective
The companies describe a ready-to-train base model and a small training-image requirement for many sample types. That is not a universal five-image guarantee: difficult or rare defects can need different examples and independent acceptance criteria.
A manufacturing trial should freeze microscope settings, separate training and blind-test samples, and record false rejects alongside missed defects. If models are shared across workstations, teams should check calibration, illumination drift, image storage and change control. Locally processed data may aid governance, but it does not replace validation of model behavior when the product or process changes.
Original source
Original source: joint Leica Microsystems/Cognex announcement published by Leica on September 29, 2026; product characteristics are company statements.