{"id":79095,"date":"2023-04-26T17:53:12","date_gmt":"2023-04-26T15:53:12","guid":{"rendered":"https:\/\/packaging-journal.de\/?p=79095"},"modified":"2023-04-26T12:56:28","modified_gmt":"2023-04-26T10:56:28","slug":"ai-powered-technologies-packaging-industry","status":"publish","type":"post","link":"https:\/\/packaging-journal.de\/en\/ai-powered-technologien-verpackungsbranche\/","title":{"rendered":"AI-powered technologies for the packaging industry"},"content":{"rendered":"
\"\"<\/a><\/div><\/div>\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Sea Vision is focussing on AI-supported technologies in its presentation at interpack. Machine vision, traceability and process automation will play a central role. Visitors can experience the technologies in practice in D\u00fcsseldorf.\u00a0<\/strong><\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Particular attention will be paid to the new technology that is based on artificial intelligence algorithms<\/strong>which were developed together with ARGO Vision and are used in the inspection of lipsticks and the automated handling of lines. At the trade fair, customers will be able to experience these technologies in practice and learn how Shaping the pharmaceutical and cosmetic manufacturing processes of the future with AI<\/strong> can.<\/p>

In the area of traceability, the main focus is on the Track & Trace software for blister packs<\/strong>which enables the inline printing of serialised graphics on primary packaging together with complete aggregation down to pallet level.<\/p>

\"\"<\/a><\/div><\/div>

Lipsticks are demanding products in terms of production and quality control. Artificial intelligence is intended to eliminate a large proportion of potential defects. The system of Sea Vision Group<\/a> uses a semantic Segmentation of the areas of the lipstick (e.g. body, tip, neck, mechanism, etc.) to identify each possible defect pixel by pixel<\/strong>. This is achieved by categorising the areas and giving each area a name or \"label\". Each part or area of the image is classified by category and identified by a colour on the screen to provide the operator with immediate information about the areas being examined.<\/p>

The System learns to recognise an ever greater variety of increasingly complex errors<\/strong>piece by piece. Learning is done both on the basis of proprietary datasets - a mixture of real and synthetic images generated using the most advanced data augmentation and neural generation techniques - and by combining the different models and parameters observed over time.<\/p>