| Name: | Yujin Chen |
|---|---|
| Position: | Ph.D Candidate |
| E-Mail: | terencecyj@gmail.com |
| Phone: | TBD |
| Room No: | 02.07.041 |
Bio
Yujin Chen is a Ph.D. student at the Visual Computing Lab advised by Prof. Matthias Nießner. His research focuses on understanding dynamic 3D environments. He received his B.Eng and M.Sc in Geo-Information at Wuhan University. Homepage
Research Interest
3D Reconstruction, Semantic Understanding, Motion Analysis, Representation LearningPublications
2025
| PBR-SR: Mesh PBR Texture Super Resolution from 2D Image Priors |
|---|
| Yujin Chen, Yinyu Nie, Benjamin Ummenhofer, Reiner Birkl, Michael Paulitsch, Matthias Nießner |
| NeurIPS 2025 |
| PBR-SR is a zero-shot approach for super-resolving physically based rendering (PBR) textures, using a pretrained super-resolution model and iterative optimization with differentiable rendering. The method refines high-resolution textures by minimizing deviations between super-resolution priors and multi-view renderings, while enforcing identity constraints to preserve fidelity to the low-resolution input. It requires no additional training or data, and consistently produces high-quality PBR textures for both artist-designed and AI-generated meshes. |
| [video][bibtex][project page] |
2024
| Mesh2NeRF: Direct Mesh Supervision for Neural Radiance Field Representation and Generation |
|---|
| Yujin Chen, Yinyu Nie, Benjamin Ummenhofer, Reiner Birkl, Michael Paulitsch, Matthias Müller, Matthias Nießner |
| ECCV 2024 |
| Mesh2NeRF is a method for extracting ground truth radiance fields directly from 3D textured meshes by incorporating mesh geometry, texture, and environment lighting information. Mesh2NeRF serves as direct 3D supervision for neural radiance fields, leveraging mesh data for improving novel view synthesis performance. Mesh2NeRF can function as supervision for generative models during training on mesh collections. |
| [video][bibtex][project page] |
2022
| 4DContrast: Contrastive Learning with Dynamic Correspondences for 3D Scene Understanding |
|---|
| Yujin Chen, Matthias Nießner, Angela Dai |
| ECCV 2022 |
| We present a new approach to instill 4D dynamic object priors into learned 3D representations by unsupervised pre-training. We propose a new data augmentation scheme leveraging synthetic 3D shapes moving in static 3D environments, and employ contrastive learning under 3D-4D constraints that encode 4D invariances into the learned 3D representations. Experiments demonstrate that our unsupervised representation learning results in improvement in downstream 3D semantic segmentation, object detection, and instance segmentation tasks. |
| [video][bibtex][project page] |