GRASP Seminar: Yinghao Xu, Stanford University, “Large Reconstruction Model for Efficient 3D Reconstruction and Generation”
May 24 at 1:00 PM - 2:00 PM
*This seminar will be held in-person in Wu and Chen as well as virtually via Zoom.
ABSTRACT
Over the past year, the large language model has achieved significant milestones, approaching human-like intelligence across various domains. However, there has been limited investigation into large-scale 3D reconstruction in the literature. In this talk, I will primarily focus on our recent advancements in large-scale 3D reconstruction.
I will start with an introduction to the basics of the Large-scale Reconstruction Model (LRM), aiming to develop a robust and highly generalizable 3D reconstruction system utilizing high-quality 3D data. I will also explain how LRM can be used to efficiently perform high-quality text-to-3D and image-to-3D generation tasks, such as Instant3D and DMV3D.Finally, I will highlight our recent work, specifically our progress in large-scale 3D reconstruction using Gaussian Splatting (GRM). As a large-scale reconstructor, GRM can reconstruct a 3D asset from sparse-view images in about 0.1 seconds. Moreover, GRM shows promising potential in generative tasks, such as text-to-3D and image-to-3D, through its integration with existing multi-view diffusion models.
Yinghao Xu
Stanford University
Yinghao Xu is a postdoctoral researcher at the Stanford Computational Imaging Lab, Stanford University, where he works under the guidance of Professor Gordon Wetzstein. Previously, he was a Ph.D. student at the Multimedia Lab (MMLab) at The Chinese University of Hong Kong. He has a deep interest in the intersection of Computer Graphics and Computer Vision. His current research focuses on generative models and neural rendering, particularly in the area of 3D generative models. Many of his papers have been accepted for oral presentation at top conferences such as CVPR, ECCV, and ICLR. One of his papers was nominated as a best paper candidate at CVPR 2020.