| CARVIEW |
Biography
I am Xingguang (Kevin) Yan, a Ph.D. student at the GrUVi lab of SFU, supervised by Prof. Angel X. Chang. Previously, I obtained my master’s degree in the Visual Computing Center (VCC) at Shenzhen University under the supervision of Prof. Hui Huang. I also worked with Prof. Daniel Cohen-Or and Prof. Niloy J. Mitra.
My research is focused on how to efficiently represent, process, and generate structured 3D data, spanning the fields of computer graphics, computer vision, and machine learning. Recent advances of deep unsupervised learning methods (such as diffusion models, autoregressive models and GANs) have unlocked remarkable success in generating text and images. I am interested in harnessing these techniques' potential to synthesize high-quality structured 3D objects and scenes, which also require special and more efficient 3D representations.
- 3D reconstruction & shape completion
- Deep generative models
- Geometry Processing
-
Ph.D. in Computer Science, starting from 2022
Simon Fraser University
-
Master in Engineering, 2021
Shenzhen University
-
BSc in Information and Computing Science, 2018
Fujian Normal University
Publications
We propose Object Images, a 2D representation that enables realistic 3D shape generation with UV maps using image-based models like Diffusion Transformers.
A transformer-based network that produces a distribution of object completions, conditioned on incomplete, and possibly noisy, point clouds.
A deep learning-based approach which simultaneously infers movable parts and hallucinates their motions from a single, un-segmented, and possibly partial, 3D point cloud shape.
Misc
I like running and have finished 1 full-Marathon, 7 half-Marathon and 1 triathlon.