LOGML - Maks Ovsjanikov: Robust learning-based methods for shape correspondence

In this talk I will describe several recent works aimed at developing accurate and robust methods for non-rigid 3D shape matching and comparison. I will first describe several ways to model this problem, including supervised, unsupervised and weakly supervised training losses. In addition, I will highlight several recent architectures that are well adapted to computing dense correspondences across a variety of settings. My ultimate goal will be to show that these techniques are becoming remarkably robust and universally applicable and useful.

Maks Ovsjanikov is a professor of computer science at Ecole Polytechnique. Prior to this, he worked for Google in their Image Search team in Mountain View. In 2011, he graduated from Stanford University with a Ph.D. in Computational Mathematics (from the ICME department), having done work on shape analysis in the geometric computing lab headed by Prof. Leonidas Guibas. His research interests are in geometric data analysis, and especially in the analysis and processing of deformable 3D shapes, with an emphasis on Deep Learning for non-rigid shape comparison and processing. Over the course of his career, Prof. Ovjsanikov introduced several key concepts in shape analysis, including the Heat Kernel Signature (cited over 1400 times, led to a Wikipedia article), algorithms for isometric shape matching, and Functional Maps. His research is the object of several international patents widely adopted in the industry. He received the Eurographics Young Researcher Award in 2014, and a Bronze medal from the CNRS in 2018.

Видео LOGML - Maks Ovsjanikov: Robust learning-based methods for shape correspondence канала LOGML Summer School