Haoyu Wu

Hi, I am Haoyu Wu (吴浩宇), a Ph.D. candidate at Stony Brook University, supervised by Prof. Dimitris Samaras. I received my bachelor's degree from the Chinese University of Hong Kong, Shenzhen.

My research focuses on computer vision and deep learning, and I have a deep passion for AI-powered content creation. My expertise includes image, 3D, and video generation, diffusion models, and neural rendering.

Email / CV / Google Scholar

Research

	One Attention, One Scale: Phase-Aligned Rotary Positional Embeddings for Mixed-Resolution Diffusion Transformer Haoyu Wu, Jingyi Xu, Qiaomu Miao, Dimitris Samaras, Hieu Le In submission, 2025 project page Analyzed RoPE interpolation failures in mixed-resolution DiTs and developed a phase-coherent attention mechanism for efficient, high-fidelity generation.
	Importance-Based Token Merging for Efficient Image and Video Generation Haoyu Wu, Jingyi Xu, Hieu Le, Dimitris Samaras ICCV, 2025 (Oral, top 1%) project page We propose an importance-based token merging method for efficient image and video diffusion model inference.
	Direct and Explicit 3D Generation from a Single Image Haoyu Wu, Meher Gitika Karumuri, Chuhang Zou, Seungbae Bang, Yuelong Li, Dimitris Samaras, Sunil Hadap 3DV, 2025 project page We propose a novel image-to-3D framework using multi-view diffusion, which scales well to high resolution by generating explicit surface geometry and texture.
	MLI-NeRF: Multi-Light Intrinsic-Aware Neural Radiance Fields Yixiong Yang, Shilin Hu, Haoyu Wu, Ramon Baldrich, Dimitris Samaras, Maria Vanrell 3DV, 2025 code / arXiv We generate pseudo-label images for reflectance and shading to guide intrinsic image decomposition for neural radiance fields.
	S-VolSDF: Sparse Multi-View Stereo Regularization of Neural Implicit Surfaces Haoyu Wu, Alexandros Graikos, Dimitris Samaras ICCV, 2023 project page Physically based neural rendering, combined with multi-view stereo, enables view synthesis from only 3 input images.
	HighEr-Resolution Network for Image Demosaicing and Enhancing Kangfu Mei, Juncheng Li, Jiajie Zhang, Haoyu Wu, Jie Li, Rui Huang ICCV Workshop, 2019 code / arXiv Winner of AIM2019: Mobile Raw to DSLR RGB Image Mapping Challenge (ICCV Workshop).

The website template was borrowed from Jon Barron.