Vincent Lepetit: "Understanding images"

CDG: Congratulations on successfully completing the CD Laboratory! Could you briefly describe what you researched on in the Laboratory?
Lepetit: We developed methods for understanding images and their 3D content. More exactly, our methods are able to not only recognize objects in images but also predict their 3D shapes and locations in space. We focused on methods that require less data than concurrent work to make them more practical.

CDG: What would you describe as the most important successes of your CD Laboratory (such as publications, patents or careers of staff members or you as Head of Laboratory)?
Lepetit: We managed to publish many papers in the most selective conferences of our field, file several patents, and provide our community with several datasets that have been reused by many research groups. I am very proud that all the students could find research positions in the industry after their PhD. For me, the project was a springboard to an ERC Advanced grant.

CDG: What do you particularly appreciate about the CD Laboratory funding model?
Lepetit: The greatest thing is probably that we could focus on the same problems for a long time, which on the long term was very productive scientifically. Having an commercial partner also kept us in touch with reality and made us focus on real problems.