Neural material generation and inverse rendering for graphics assets.
I work on computer graphics, neural rendering, and generative models, with a focus on extracting and synthesizing physically meaningful material appearance from images and videos. My recent research studies how generative video models can be used as material priors for 3D asset creation.
News
Recent updates
Launched the VideoNeuMat project page and personal academic homepage.
VideoNeuMat appears at ACM SIGGRAPH 2026.
PBR-Inspired Controllable Diffusion for Image Generation appears in Computer Graphics Forum, Eurographics 2026.
Featured Project
ACM SIGGRAPH 2026
VideoNeuMat: Neural Material Extraction from Generative Video Models
VideoNeuMat reconstructs neural materials from generative video observations, enabling high-resolution appearance, relighting, and geometry-aware material transfer for graphics workflows.