The past decade in computer vision has witnessed the re-emergence of deep learning, and in particular convolutional neural network (CNN) techniques, allowing to learn powerful image feature representations from large collections of examples. Yet, when attempting to apply the CNN paradigm to 3D shapes and point clouds (similarity, correspondence, retrieval, etc.) one has to face fundamental differences between images and geometric objects.
Shape analysis and 3D vision pose new challenges that are non-existent in image analysis, and deep learning methods have only recently started penetrating into the 3D community. The purpose of this session is to overview the foundations and the current state of the art on learning techniques for 3D shape analysis and vision. Special focus will be put on techniques applied to non-Euclidean manifolds for tasks of shape classification, object recognition, retrieval and correspondence, presenting in a new light a broad showcase of geometric problems.
Emanuele Rodola’: Professor of CS at Sapienza University of Rome
Emanuele Rodolà is Professor of CS at Sapienza University of Rome, where he leads the GLADIA group of Geometry, Learning and Applied AI, funded by an ERC Grant and a Google Award. Previously, he held positions at USI Lugano, TU Munich, and the University of Tokyo. His research interests lie at the intersection of geometry processing, graph / geometric deep learning, applied AI and computer vision, and has published more than 100 papers in these areas.