REPARTI Webinar: Understanding the world behind the image
Jean-François Lalonde
Laboratoire de Vision et Systèmes Numériques (LVSN)
Dép. de génie électrique et de génie informatique, Université Laval
March 29, 2021, 12:00 – 1:00 p.m.
Abstract
Images are formed through a series of interactions between light, surfaces in the scene (according to their geometry and reflectance) and, ultimately, the camera. Accurate physical models of these interactions exist, but have seen limited applicability in real-world conditions, outside the lab. Because of this, a large fraction of computer vision research treats images as 2D pixel arrays without regard to how they were formed.
In this talk, I instead advocate for the idea of reasoning about the real world behind the image and explicitly consider these light-geometry-camera interactions. To do so, we propose algorithms that understand the 3D geometry, lighting, surface reflectance, and even the camera itself from images. The key idea is to combine physics-based models and machine learning techniques in order to better model, understand, and recreate the richness of our visual world.
The presentation will be given in French and the slides will be in English.
Jean-François Lalonde, Ph.D., is an Associate Professor in the Faculty of Science and Engineering at Université Laval, in the Department of Electrical and Computer Engineering. He is a member of the Institute Intelligence and Data (IID), the Big Data Research Center (CRDM), and the Research Center on Vision, Robotics and Machine Intelligence (CeRVIM) at Université Laval. Previously, he was a Post-Doctoral Associate at Disney Research, Pittsburgh. He received a Ph.D. in Robotics from Carnegie Mellon University in 2011. His thesis, titled Understanding and Recreating Appearance under Natural Illumination, won the CMU School of Computer Science Distinguished Dissertation Award. His research interests lie at the intersection of computer vision, computer graphics, and machine learning. In particular, he is interested in exploring how physics-based models and data-driven machine learning techniques can be unified to better understand, model, interpret, and recreate the richness of our visual world. His group has captured and published the largest datasets of indoor and outdoor high dynamic range illumination images, freely available for research. He is actively involved in bringing research ideas to commercial products, as demonstrated by his several patents, technology transfers with large companies such as Adobe and Facebook, and involvement with startups including Geomagical Labs (San Francisco, acquired by IKEA) and TandemLaunch (Montreal).
Zoom Meeting
To obtain the Zoom meeting web link, please contact:
Annette.Schwerdtfeger@gel.ulaval.ca
