DUETT: First Lecture in the “Building Blocks of Visual AI” Lecture Series

On August 10, 2026, another knowledge transfer event took place as part of the DUETT knowledge and technology transfer project. In his presentation titled “Behind the Scenes of Visual AI—Building Blocks of Current AI Models in Image Recognition,” Oliver Kempf, the lead researcher for Use Case 1, provided insight into the fundamental methods and mechanisms of modern AI models for image processing.

The presentation marked the start of the lecture series “Building Blocks of Visual AI,” which, as part of Use Case 1, will explore further aspects of visual artificial intelligence and its application in plastics processing. Use Case 1 focuses on the inline detection and classification of visual defects in injection-molded parts.

The presentation explained the key building blocks of visual AI and how they interact in the automated detection and processing of image information. The event offered participants the opportunity to learn about the fundamental concepts of current AI models and to better understand their potential applications in industrial settings.

A total of 49 people participated in the online event, including members of the project steering committee from companies, chambers of commerce, and networks; interested staff from the Department of Plastics Engineering; and the DUETT project team, consisting of Oliver Kempf, Marco Klute, and Alisa Bangert.

Further information about the project, as well as the slides and a recording of the online presentation, are available on the DUETT project website.

A separate series of presentations is also planned for Use Case 2. This series is scheduled to begin in late 2026 under the title “AI in Recyclate Processing” with an initial event on the topic of“Characterization of Recyclates” and will address various aspects of data-driven recyclate processing as the project progresses. The exact date will be announced at a later time and communicated well in advance.