Neural Mesh Processing

Vladimir G. Kim

Senior Research Scientist at Adobe Research

Apr 06, 2021 (Tue), 9:00 a.m. KST

Guest Lecture at CS492(H) Machine Learning for 3D Data (Spring 2021), KAIST.

Host: Minhyuk Sung

Add to Google Calendar

Zoom Webinar Link

Abstract

Creating high-quality 3D models requires significant expertise and artistic skill. The overarching objective of my work is to make generation and manipulation of 3D surfaces more accessible and intuitive for non-experts by developing novel machine learning techniques and representations for 3D geometry. I will present several techniques that were inspired by classical geometry processing tools, such as surface parameterization via atlases, mesh deformation, and subdivision surfaces. By re-envisioning these classical techniques as novel modules and layers in neural network architectures, we can use them to solve complex problems that require optimizing over collections of shapes or learning non-trivial task-specific priors.

Bio

Vladimir G. Kim is a Senior Research Scientist at Adobe Research, Seattle. He works on geometry analysis algorithms at the intersection of graphics, vision, and machine learning, enabling novel interfaces for creative tasks. His recent research focuses on making it easier to understand, model, manipulate, and process geometric data such as 3D meshes and point clouds.