I’m currently an undergraduate student studying in Information Science Advanced Class, school of Electronic Information & Communication, HUST. I have a great interest in many aspects in Artificial Intelligence, especially computer vision related tasks.

Now I am a visiting student at VlLab@UCMerced, supervised by Prof. Ming-Hsuan Yang, also working closely with Dr. Lu Qi and Dr. Xueting Li. Previously, I am very fortunate to be supervised by Prof. Xinggang Wang@HUSTVL, Prof. Meng Tang@mengtang lab, UCMerced.

External collaborations are welcomed, please contact me via email.

Also if you are interested in a chat, please send me an email directly. Always welcome!

Contact: colezwhyzz [at] gmail [dot] com

• Expected to be seeking for Ph.D. positions in 2025 Fall!

Research Interest

• Artificial Intelligence, Computer Vision, Multi-modality

• data-efficient learning, General visual intelligence

• Generative-related, 3D/4D generation, AIGC

• To be explored…

Always open for interesting topics!

News

• 2024/10/22: Layout-Your-3D now availlable on arxiv.

• 2024/7/21: Our work WeakSAM is accepted to ACM Multimedia 2024!

• 2024/6/8: Starting my internship at UCMerced and working on generative tasks. Good luck!

• 2024/3/30: I will join VLLab@UCMerced as a visiting student, supervised by Prof. Ming-Husan Yang.

• 2024/2/22: Our work WeakSAM is now available on Arxiv, please check the full paper and refer to our open-sourced codes for more details.

Education

• 2021 Sep. - 2025 Jun.(expected):

  B.E. in Electronic Information Engineering

  Huazhong University of Science & Technology

Research & Working experiences

• 2024 Jun. - Present.:
  • UCMerced VLLab – Visiting Student
  • The University of California, Merced
  • Content:
    • 3D/4D generation
    • Reconstruction
  • Supervisor: Ming-Hsuan Yang

• 2022 Oct. - 2023 Dec.:
  • HUSTVL – Research Intern
  • Huazhong Univeristy of Science & Technology
  • Content:
    • General Visual Intelligence
    • Feature Enginnering
    • Vision-Language
    • Unsupervised Learning
  • Supervisor: Xinggang Wang

• 2023 Dec. - 2024 Mar.:
  • Mengtang Lab – Research Intern
  • The Univerisity of California, Merced
  • Content:
    • Diffusion for perception
    • Efficient Network Structural design
    • Image Post refinement for generative models
  • Supervisor: Meng Tang

Publications

Layout-your-3D: Controllable and Precise 3D Generation with 2D Blueprint

Junwei Zhou, Xueting Li, Lu Qi, Ming-Hsuan Yang

WeakSAM: Segment Anything Meets Weakly-supervised Instance-level Recognition

Lianghui Zhu*, Junwei Zhou*, Yan Liu, Hao Xin, Wenyu Liu, Xinggang Wang

* stands for equal contribution

Service

Reviewer for ICLR2025.