For my work on unsupervised learning for graph matching I received, in 2014, the “Grigore Moisil” Prize, the most prestigious award in Mathematics given by the Romanian Academy. Currently, I lead research projects on various computer vision and deep learning problems. They span a diverse but well-connected directions of research, which aim in complementary ways to achieve, with as little supervision as possible, a coherent semantic understanding of the visual world in space and time. Such research areas include unsupervised learning, optimization methods for graphs and deep neural networks, object tracking, semantic segmentation, vision and learning for Earth Observation, activity recognition in video, as well as connecting vision and language.

I tend to look for a balance between theory and practice, therefore I love to get involved in projects that take full advantage of intelligent algorithms as well as real world data. Therefore, some of my projects involve developing vison and learning systems for drones, self-driving cars or medical applications as well as the wood industry. As it can be seen on this site, some of these projects are done in collaboration with top IT companies.