Aivaro AI Feature Detection identifies, classifies and measures defects inside Leica’s Enersight inspection environment without requiring cloud processing.
Leica Microsystems and Cognex have introduced Aivaro AI Feature Detection, an artificial-intelligence software module for routine industrial inspection and quality-control applications.
The module is integrated into Leica’s Enersight software platform and uses machine-vision technology from Cognex to identify, classify and measure defects or other features of interest in microscope images.
Aivaro is intended to improve the consistency of inspection decisions across different operators, workstations and production lines. Stored microscope settings and imaging parameters can be reproduced, helping manufacturers apply the same inspection process at several stations or facilities.
The software also generates statistical information, including minimum, maximum and average measurements. This allows detected features to be incorporated into quality-control reports rather than serving only as visual classifications.
Leica says the supplied base model can be trained for an application using between five and ten images. The actual number required for dependable operation will depend on the variation in the inspected parts, imaging conditions and range of defects that the model must recognise.
Processing runs locally on standard hardware at the production site. The system therefore does not require cloud connectivity, which may be important for factories handling proprietary designs, regulated products or inspection data that cannot be transferred outside the facility.
Potential applications include electronics manufacturing, battery production, transportation equipment, medical devices and polymer processing. In these environments, conventional manual microscopy can produce differences between operators, particularly when defect boundaries or acceptance criteria are difficult to interpret consistently.
The new module leaves final inspection decisions under the manufacturer’s control while automating repetitive detection and measurement work. Its integration into an existing microscope-software platform is intended to make AI-assisted inspection accessible without requiring users to develop a separate machine-vision system.
Image credit: Leica Microsystems / Cognex
