Chuhan Zhang (张楚晗)

I am a researcher working on multimodal AI, with interests spanning video understanding, dynamic 3D scene reconstruction, and automatic evaluation pipelines for generative models. I am particularly keen on building models that are simple in design yet achieve strong performance. My recent work focuses on video spatial understanding and situated awareness in Gemini.

I completed my PhD at the Visual Geometry Group (VGG), University of Oxford, advised by Andrew Zisserman. Prior to my PhD, I obtained my MEng in Engineering Science at Exeter College, University of Oxford. Please feel free to reach out by email if you'd like to discuss research or collaborate.

Selected Publications

• Efficiently Reconstructing Dynamic Scenes One D4RT at a Time
  Chuhan Zhang, Guillaume Le Moing, Skanda Koppula, Ignacio Rocco, …, Andrew Zisserman, Junlin Zhang, Mehdi S. M. Sajjadi
  IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026 Oral Award Candidate
  PDF arXiv Project Blog
• Scaling 4D Representations
  João Carreira, Dilara Gokay, Michael King, Chuhan Zhang, Ignacio Rocco, …, Andrew Zisserman
  arXiv preprint, 2024
  arXiv Code
• Revisiting Text-to-Image Evaluation with Gecko: On Metrics, Prompts, and Human Ratings
  Olivia Wiles^*, Chuhan Zhang^*, Isabela Albuquerque^*, …, Aida Nematzadeh
  ^*Equal contribution
  International Conference on Learning Representations (ICLR), 2025 Spotlight
  arXiv OpenReview
• Helping Hands: An Object-Aware Ego-Centric Video Recognition Model
  Chuhan Zhang, Ankush Gupta, Andrew Zisserman
  International Conference on Computer Vision (ICCV), 2023
  PDF Code
• Is an Object-Centric Video Representation Beneficial for Transfer?
  Chuhan Zhang, Ankush Gupta, Andrew Zisserman
  Asian Conference on Computer Vision (ACCV), 2022
  PDF arXiv
• Temporal Query Networks for Fine-grained Video Understanding
  Chuhan Zhang, Ankush Gupta, Andrew Zisserman
  IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021 Oral
  PDF arXiv Code Project
• Adaptive Text Recognition through Visual Matching
  Chuhan Zhang, Ankush Gupta, Andrew Zisserman
  European Conference on Computer Vision (ECCV), 2020
  PDF Project Code