Stevanato Group is now using artificial intelligence (AI) in visual inspection to address the problem of false rejects. The technology could potentially be used even more widely to classify different categories of rejects or identify foreign particles, providing critical information to improve the production process.
AI enables systems to correlate the actual image of the product under inspection with the content of a huge library of similar images for accepted and rejected products and assign the likelihood of belonging to one of the categories in the library. The pharma industry has been actively studying this technology in the last two years and Stevanato Group is among the forerunners.
One of the challenges in implementing AI for visual inspection is the limited availability of pharmaceutical image datasets compared to those available for everyday objects such as cars or animals, which have accelerated the development of technologies such as face recognition, image captioning, and autonomous driving. Additionally, inspection images are often part of customers' process data, requiring strict guarantees of data integrity, security, and controlled accessibility for validation purposes by regulatory agencies such as the FDA and EMA.
With standard visual technology, operators use classical measurement parameters, relying on their experience and production samples. False rejects are the inevitable trade-off resulting from the fact that only a limited set of samples is available. In some cases, good product may be rejected to maintain a conservative approach.
AI enables the automatic analysis of hundreds of thousands of images. This provides a richer representation of defects and a stronger knowledge base from which to start parameterization. As a result, false reject rates can be significantly reduced without compromising detection performance.
Stevanato Group is actively supporting a customer in implementing a prototype system in-line, while other customers are comparing the results of AI in addressing specific visual inspection challenges related to container or drug products. The company is also experimenting with AI in its own production lines to reduce false rejects.
