Ray (Rui) Zhang 张锐

photo_25.jpg

Toyota Research Institute

4440 El Camino Real

Los Altos, CA 94022

I’m currently a Research Scientist at Toyota Research Institute, where I develop realtime perception algorithms for autonomous robots. My research interest lies in the theory and applications of robot perception, computer vision, and geometric deep learning methods.

Previously, I completed my Robotics PhD at Computational Autonomy and Robotics Laboratory (CURLY) and the Perceptual Robotics Laboratory (PeRL) at University of Michigan, Ann Arbor, adviced by Professor Maani Ghaffari and Professor Ryan Eustice. I also had collabrations with Professor Jessy Grizzle and research internships at Amazon Lab126 and Mathworks. Additionally, I served as a student investor in the space of Robotics/AI at Wolverine Venture Fund.

news

Jun 1, 2026 One paper accepted at CVPR 26 (Highlight).
May 1, 2026 One paper accepted at ECC 26
Dec 1, 2025 One paper accepted at TPAMI.

selected publications

Full list on Google Scholar

  1. CVPR
    Generalized-CVO: Fast and Correspondence-Free Local Point Cloud Registration with Second Order Riemannian Optimization
    Zhang, Ray, Greiff, Marcus, Lew, Thomas,  and Subosits, John
    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Jun 2026
  2. TPAMI
    RKHS-BA: A Robust Correspondence-Free Multi-View Bundle Adjustment Framework for Semantic Point Clouds
    Zhang, Ray, Song, Jingwei, Gao, Xiang, Wu, Junzhe, Liu, Tiany, Zhang, Jinyuan, Eustice, Ryan,  and Ghaffari, Maani
    IEEE Transactions on Pattern Analysis and Machine Intelligence Jun 2025
  3. ECCV
    Correspondence-Free SE (3) Point Cloud Registration in RKHS via Unsupervised Equivariant Learning
    Zhang, Ray, Zhou, Zheming, Sun, Min, Ghasemalizadeh, Omid, Kuo, Cheng-Hao, Eustice, Ryan, Ghaffari, Maani,  and Sen, Arnie
    Proceedings of the European conference on computer vision Jun 2024

projects

  • Generalized-CVO

    Fast and correspondence-free point cloud registration in RKHS with second-order Riemannian optimization

code