Cooperating with companies is part of the very DNA of the University of Luxembourg’s Interdisciplinary Centre for Security, Reliability and Trust (SnT). Its extensive partnership programme includes major Luxembourg companies such as IEE, POST and SES as well as highly innovative smaller companies and start-ups. “Almost everything we do is in collaboration with industry,” says Djamila Aouada, who heads the centre’s Computer Vision, Imaging and Machine Intelligence Research Group.

Computer vision expertise

Professor Djamila Aouada heads the Computer Vision research group at the University of LuxembourgProfessor Aouada’s group carries out research on computer vision, image processing, image analysis, visual data understanding and machine learning. “My PhD – defended in 2009 – was on 3D object recognition and 3D shape modelling,” she explains. “I continue developing this research with my team, extending to more complex 3D shapes such as human faces and articulated bodies, and recently to inferring computer-aided design, or CAD models, from 3D scans.”

My strategy is to launch research topics that are motivated by industry, and use the insights gained to perform more fundamental research.

She gained her first experience of public-private research partnerships when joining SnT after her doctoral studies. “I immediately initiated a collaboration with sensing solutions specialist IEE. The topic was still 3D data modelling, but the angle and requirements were completely different from what I had done before. I learnt so much from that project.”

Since then, research partnerships have been at the centre of Professor Aouada’s work. “My strategy is to launch research topics that are motivated by industry, and use the insights gained to perform more fundamental research.”

Needs-oriented research partnerships

The research group collaborates with a wide range of partners with varying needs and on different computer vision and machine learning tasks. “One example is Artec 3D, the world leader in handheld 3D scanners. Like us, they are real experts in 3D computer vision. We cooperate closely on techniques and algorithms, and the company benefits from our partnership to push fundamental research further.”

Almost everything we do is in collaboration with industry.

Another, different example is UK space start-up LMO, which was attracted to Luxembourg by its dynamic space ecosystem as well as by its research expertise. “The LMO team has advanced knowledge in the field of electronics, but needs computer vision skills to develop its autonomous space systems. We provide them with our expertise and help them build their technology. At the same time, we learn about the needs of the space sector from them.”

Other companies have worked with the team on topics such as deep fake detection, multi-modal data fusion for autonomous driving, industrial quality control and the detection of anomalies in financial transaction data.

Knowledge and relevance – the joint benefits

Professor Aouada points out that public-private research partnerships are both challenging and rewarding. “Working with partners is demanding: they have a business to run and expect results. In many ways, it is easier just to do fundamental research and hope that it will benefit someone, some day. The big advantage with partnerships is that we are constantly learning and seeing the direct impact of our work in very different areas. Each partner represents a unique story.”

Photo: Luxinnovation/Luc Deflorenne