Daxuan (达萱) Ren (任)
Ph.D Student at Nanyang Technological University, Singapore
- NTU School of Computer Science and Engineering (SCSE)
- Algorithm Researcher @ SenseTime Research
- Sensetime-NTU Industrial Post-Graduate Program (IPP)
- Email: daxuan001 [at] e.ntu.edu.sg
I am currently a Ph.D student at the School of Computer Science and Engineering (SCSE) at Nanyang Technological University (NTU), Singapore. I am under the supervision of Prof. Zheng Jianmin (NTU) and Prof. Cai Jianfei (Monash University). I’m under the Indistrial Post-Graduate Program (IPP), which is a joint program between SenseTime and NTU initiated by Singapore Economic Development Board (EDB).
I received my bachelor degree in Computer Science from NTU in 2018. During my undergraduate study, I was interested in High-Performance-Computing(HPC) and privileged to represented NTU to compete with other top universities around the globe in the International Supercomputer Conference (ISC) Student Cluster Competetion in Frankfurt, Germany, and ACM/IEEE Supercomputer Conference (SC) Student Cluster Competetion in Salt Lake City, USA.
Prior to my Ph.D, I was a software engineer in Autodesk working on large-scale 3D Reconstruction Engine (Autodesk ReCap Photo) including Structure-from-Motion (SFM), Multi-View-Stereo (MVS) and Surface Reconstruction from 2017 to 2020.
My research interests and past experiences include Multiview Reconstruction(camera calibration, pointcloud, mesh and texture reconstruction from multiview RGB or Kinect-RGBD), Shape Parsing (turing rastered 3D data into parametric CAD shapes), physically based rendering and differentiable rendering (via ray tracing or volumetric rendering but not rasterzation) and 3D computer vision and graphics in general.
News
Sep 20, 2022 |
Three Papers : “ExtrudeNet: Unsupervised Inverse Sketch and Extrude for Shape Parsing”, “HuMMan: Multi-Modal 4D Human Dataset for Versatile Sensing and Modeling” and “Monocular 3D Object Reconstruction with GAN Inversion” was accepted at ECCV 22. |
---|---|
Dec 3, 2021 | We have released MMHuman3D: 3D Human Parametric Model Toolbox and Benchmark. |
Jul 21, 2021 | Our paper “CSG-Stump: A Learning Friendly CSG-Like Representation for Interpretable Shape Parsing” was accepted at ICCV 21 |
Publications
- ExtrudeNet: Unsupervised Inverse Sketch and Extrude for Shape ParsingIn European Conference on Computer Vision 2022
- Monocular 3D Object Reconstruction with GAN InversionIn European Conference on Computer Vision 2022
- HuMMan: Multi-Modal 4D Human Dataset for Versatile Sensing and ModelingIn European Conference on Computer Vision 2022
- Playing for 3D human recoveryarXiv preprint arXiv:2110.07588 2021
- CSG-Stump: A Learning Friendly CSG-Like Representation for Interpretable Shape ParsingIn International Conference on Computer Vision 2021
- Messytable: Instance association in multiple camera viewsIn European Conference on Computer Vision 2020