Supercharging visual inspection with artificial intelligence

Criterion AI leverages the power of AI to significantly decrease false rejection rates and to increase detection rates for pharmaceutical products and medical devices.

GxP Compliant Platform to Leverage AI for Visual Inspection

Increase output of visual inspection lines and reduce maintenance

Traditional rules-based computer vision solutions have been deployed widely across the pharmaceutical and medical device industries over the past couple of decades. While they have enabled a high degree of automation in manufacturing and inspection processes, they often have the downside of providing less-than-optimal rejection and detection rates.

AI-based algorithms significantly outperform traditional algorithms by being based on data-driven training approaches rather than arbitrary rules and thresholds.

Increased patient safety by improving detection of known/unknown defects

Criterion AI’s state-of-the-art AI algorithms for classification of images provide much higher detection rates of known defects than traditional rules-based computer vision algorithms.

This is largely thanks to the fact the AI algorithms can be trained on large amounts of images of real products obtained from production, containing natural process variation.

In addition, Criterion AI makes use of unsupervised algorithms that can be trained to catch unknown defects by establishing a baseline of what good products look like. These algorithms ensure that defects that were not known at the time of training can be catched in production when they occur.

Strengthen the feedback loop across production processes

Predictions made by Criterion AI’s algorithms running on visual inspection lines are being stored in the cloud and can be visualized through interactive real-time dashboards. These make it easy to communicate potential quality issues and can serve as the basis for much more constructive dialogs on how to improve upstream production processes.

Moreover, predictions can be correlated with other relevant production data to identify root causes for quality issues in a statistical manner and to potentially control production parameters in a closed loop. This way, intelligent solutions can be built to continuously maintain high levels of quality.

Interested in learning more about how to improve visual inspection with AI?

This website collects usage data to help us improve your experience of the site. Click here to accept or here to opt out.