Earlier projects
Selected projects from my undergraduate studies in computer graphics, physics-based simulation, and computer vision. My current research is described on the home page.
Feature-Grid Based Texture Synthesis and Transfer
A method for texture synthesis and transfer based on a feature-grid neural implicit representation of an arbitrary object, which allows such an object to be transferred as a texture onto an arbitrary target surface. Because the representation is a 2D feature grid, standard computer vision and 2D geometry processing techniques can be applied directly to 3D content. Developed at the Dynamic Graphics Project (DGP) Lab.
An Approach Toward Human Musculoskeletal System Simulation
The human musculoskeletal system is affected by many factors such as muscle strength, bone stiffness, and mass, and is hard to simulate as a whole. This project builds a basic unit of bone and muscle from simple objects (mass points, rods, and springs), simulates it, and applies machine learning for fast approximation of the system state. An undergraduate research project (CSC494).
Accurate Large Scale Ferrofluids Simulation
Ferrofluids form complex spikes under a magnetic field, a shape that is hard for artists to create by hand. This project implements the ferrofluid simulation of "On the Accurate Large-scale Simulation of Ferrofluids" by Huang et al. A team course project for CSC417.
Anime Art Style Picture Super-Resolution Network
Super-resolution (denoising and upscaling) for images in the anime art style. Instead of a large deep convolutional network, this project uses a much smaller network with fewer layers and trainable parameters, and achieves comparable results with higher efficiency.
Biharmonic Distance
Measuring distances between pairs of points on a 3D surface is a fundamental problem in computer graphics and geometry processing. This project implements the biharmonic distance of Lipman et al.