Keywords: Computer Graphics, Computer Vision, Neural Rendering, Physically Based Simulation, Deep Learning.
Interests: During my Masters, my research was devoted to physically based simulations; now, I explore neural rendering techniques (NeRF/3DGS etc.) at TUM. I am interested in a wide range of topics under the topic of Computer Graphics/Vision.
[07/2025] I am currently looking for a new Ph.D. position due to VISA issues!
Publications
A Second-Order Explicit Pressure Projection Method for Eulerian Fluid Simulation Junwei Jiang,
Xiangda Shen,
Yuning Gong,
Zeng Fan,
Yanli Liu,
Guanyu Xing,
Xiaohua Ren*,
Yanci Zhang*
ACMSIGGRAPH/ Eurographics Symposium on Computer Animation 2022/ Computer Graphics Forum pdf |
abstract |
video |
code |
bibtex
In this paper, we propose a novel second-order explicit midpoint method to address the issue of energy loss and vorticity dissipation in Eulerian fluid simulation. The basic idea is to explicitly compute the pressure gradient at the middle time of each time step and apply it to the velocity field after advection. Theoretically, our solver can achieve higher accuracy than the first-order solvers at similar computational cost. On the other hand, our method is twice and even faster than the implicit second-order solvers at the cost of a small loss of accuracy. We have carried out a large number of 2D, 3D and numerical experiments to verify the effectiveness and availability of our algorithm.
Parallel Real-Time Bone Driven Secondary Deformation Based on Clique Coloring
Yusong Fu,
Junwei Jiang,
Yina Lv,
Yanci Zhang*
China Computer Aided Design and Computer Graphics (CAD/CG 2021) abstract |
simple video
We propose a new clique based weighted mixed coloring decoupling scheme to reduce the error caused by par-allel Gauss Seidel iterations in dense systems such as bone driven secondary deformations. The coloring algo-rithm mainly focuses on the clique structure in the graph, that is, the global priority is given to large clique structure, and the local consideration is given to load balancing. Compared with previous work, our scheme has better decoupling effect for dense physical systems with a large number of coupling, which can make the GPU parallel solver have higher computational performance and better accuracy.
Projects
Large Scale Real-Time Fluid Simulation video
A NetEase pre-research project, which uses mixed Lagrangian particles, Eulerian grids and shallow water equation height fields to achieve a compromise between performance and accuracy.
Real-Time Remote Assistance System Based on Hololens
Develop a remote assistance system similar to Microsoft Dynamics 365. With the help of Hololens, guidance experts in different places can share the screen with front-line engineers remotely and in real time, and give instructions on the screen to achieve the function of remote assistance.
Real-Time Dynamic Diffuse Global Illumination Based on Light Field Probe
Record the light field information of the scene by light field probes. Use the specially designed algorithm to query and calculate the multiple diffuse reflections.