Go to main content

DeepReview™ for 3D AOI

DeepReview™ for 3D AOI

  • AI-powered Deep learning algorithm automatically classifies defects detected by 3D AOI
  • Fast Reduce review times by eliminating 50-100% of false calls for eligible components
  • Boost FPY Automatically prevent escapes while massively reducing false calls

Accelerate review times. Eliminate the majority of false calls. And automatically improve your defect classification models over time. DeepReview, a new AI-based Automatic Defect Classification system for 3D AOI, makes it possible to apply full test coverage to any product mix without fear of false calls.

Give review operators the focus they deserve

If a task is tedious but necessary, it may be a good candidate for AI. This is particularly true when it comes to reviewing false defects, which is both time-consuming and exhausting – distracting review operators from their primary task of identifying real defects.

Waste less time on false calls

This is why we’ve created DeepReview, a new deep learning system that trains itself to analyze new components, automatically classify defects and eliminate the vast majority of false calls. Over time, the system evolves to recognize new components in new contexts, helping operators improve their defect classification models, accelerate inspection reviews and improve their overall productivity.

High-performance ADC for high-mix production

Changing component supply chains. New boards. New layouts. In high-mix production, change is the name of the game. But the more variety that’s introduced into production, the more Automated Optical Inspection systems can struggle to stay ahead. The result is often an uphill battle against newly generated false calls in the effort to maintain full coverage.

DeepReview offers a path forward. Using dozens of layers of neural networks, it utilizes actual inspection images to build recognition of various defect sources over time — regardless of component manufacturer or imaging variations. The result is a consistent reduction in unnecessary false calls, with a corresponding reduction in review times.

Your data, your process standards

By occasionally training the DeepReview neural network with your own data, you retain full control over the quality of your process. This allows you to apply, adapt and refine your inspection models in full accordance with your own inspection standards.

The AI journey has just begun

Already embedded into MYWizard software for polarity check, DeepReview is designed to address a range of other defect sources such as solder joints, bridges, offsets and coplanarity, with new defect classification capabilities rapidly being rolled out over time.

Join us on the journey toward a new era in Automatic Defect Classification.

The DeepReview AI engine features:

  • Maximum data security and no latency thanks to direct connection with edge AI computing server
  • Powerful Dual GPU card for training of neural network
  • Full traceability with terabytes of data storage and RAID backup architecture