Syntegon Technology has successfully validated the world's first inspection machine for syringes with artificial intelligence in a customer's production line.
With the installation of a fully validated visual inspection system with artificial intelligence (AI) in a fully automated inspection machine, the company says it is breaking new ground in the industry. The aim is to establish AI in pharmaceutical inspection for various products and packaging materials in the long term.
First inspection machine with AI
Syntegon recently launched the first AI-based visual inspection system installed in a fully automated, validated customer line. The biotechnology company Amgen uses the system to reliably distinguish foreign particles from air bubbles on the rubber stopper of syringes. Previously, these were often incorrectly rejected as faulty.
"This challenging project required a great deal of commitment and expertise. In collaboration with Syntegon, we were able to realise the World's first syringe inspection machine with AI realised. This underlines our market position, both in biotechnological production and in the use of innovative technologies." Manuel Soto, senior process development engineer at Amgen
In this customer project, the AI-based image processing system from Syntegon was able to Increase particle detection rate by 70 per cent and at the same time the Reduce false rejects by 60 per cent (average value of a specific test station).
Syntegon sees this as a "milestone". In view of increasingly complex drugs, inspection in the pharmaceutical manufacturing process requires ever more sophisticated visual systems. "Especially with high-priced products, every single case of false rejects is one too many," explains Dr José Zanardi, responsible for applications and developments in visual inspection technology at Syntegon.
Inspection with AI and deep learning
AI applications enable higher detection rates and can reduce faulty rejects. Especially with difficult products such as highly viscous parenteral solutions, it is sometimes difficult to distinguish air bubbles from harmful particles. AI uses deep learning algorithmsthat are able to reliably identify recurring patterns and deviations.
"Visual applications with deep learning are increasingly available on the market. Our task was to analyse these Applications for pharmaceutical purposes adapt. This includes validation in particular," says Zanardi. Thanks to its extensive software and inspection expertise, the company was able to develop a solution that requires only minor adjustments to the existing image processing systems. Syntegon plans to implement the AI in other inspection machines for different products and packaging materials.
[infotext icon]
Bei der Entwicklung und Herstellung visueller Inspektionsmaschinen verfügt Syntegon Technology, die ehemalige Verpackungssparte ‚Bosch Packaging Technology‘ aus der Bosch-Gruppe, über mehr als 40 Jahre Erfahrung. Das Unternehmen mit Hauptsitz in Waiblingen bei Stuttgart deckt eine breite Palette an manuellen, halb- und vollautomatischen Lösungen für alle flüssigen Pharmazeutika und Packmittel ab.[/infotext]
Source: Syntegon Technology
