Hongyu Wen

I am a third year Ph.D candidate in Computer Science at Princeton University, advised by Prof. Jia Deng as a member of the Princeton Vision & Learning Lab.

Before that, I was graduated from Turing Class, EECS, Peking University, advised by Prof. Tingting Jiang. I also had the privilege of working remotely with Prof. Deepak Pathak at CMU and Ronghang Hu in FAIR.

My research interests lie in broad 3D vision, focusing on recovering geometry and scene intrinsics from visual cues and correspondences across multiple images, encompassing tasks such as depth estimation, optical flow, and 3D reconstruction. I believe that if the world can be viewed as a big simulator, then it should be possible to reverse-engineer it.

Email  /  GitHub  /  Google Scholar  

Research

Seeing and Seeing Through the Glass: Real and Synthetic Data for Multi-Layer Depth Estimation

LayeredFlow: A Real-World Benchmark for Non-Lambertian Multi-Layer Optical Flow

Infinigen Indoors: Photorealistic Indoor Scenes using Procedural Generation

Infinigen: Infinite Photorealistic Worlds using Procedural Generation

ASFormer: Transformer for Action Segmentation

Teaching