Lite Any Stereo: Efficient Zero-Shot Stereo Matching
Junpeng Jing, Weixun Luo, Ye Mao, and Krystian Mikolajczyk
Ye Mao
PhD candidate at Imperial College London working on 3D scene understanding, spatial intelligence, and multimodal large language models.
Building generalizable 3D perception systems that bridge geometry, semantics, and open-world reasoning across real-world environments.
Affiliation
Imperial College London
Focus
3D scene understanding, Spatial Intelligence, VLMs
Profile
About
Research interests, academic background, and the broader agenda guiding current work.
I am a PhD candidate in Computer Vision and Machine Learning at MatchLab, Imperial College London, supervised by Prof. Krystian Mikolajczyk and supported by the Imperial President's Scholarship .
My research focuses on developing generalizable spatial intelligence models that can represent and reason about real-world indoor environments, while transferring effectively across a wide range of 3D tasks.
Before starting my PhD, I completed an MPhil in Medical Sciences at the University of Cambridge and an MSc in Applied Machine Learning at Imperial College London. I previously studied Computer Science and Robotics at King's College London.
Current Interests
- Visual-spatial Intelligence
- 3D scene understanding
- Vision-language models
- Representation Learning
Research Direction
- Bridging geometry and semantics for unified 3D representation learning
- Learning robust 3D encoders based on point maps
- Spatial reasoning over real-world, dynamic scenes
Research Output
Publications
Selected publications spanning 3D scene understanding, stereo matching, multimodal learning, and medical imaging.
POMA-3D: The Point Map Way to 3D Scene Understanding
Ye Mao, Weixun Luo, Ranran Huang, Junpeng Jing, and Krystian Mikolajczyk
DYNAMIC: A Novel Software Implementation of a Kinetic Model of TaqMan PCR
Louis Kreitmann, Ye Mao, Ke Xu, Alison Holmes, Karen Brengel-Pesce, Laurent Drazek, and Jesus Rodriguez-Manzano
Stereo Any Video: temporally consistent stereo matching
Junpeng Jing, Weixun Luo, Ye Mao, and Krystian Mikolajczyk
Hypo3D: Exploring Hypothetical Reasoning in 3D
Ye Mao, Weixun Luo, Junpeng Jing, Anlan Qiu, and Krystian Mikolajczyk
OpenDlign: Enhancing Open-World 3D Learning with Depth-Aligned Images
Ye Mao, Junpeng Jing, and Krystian Mikolajczyk
Match-Stereo-Videos: Bidirectional Alignment for Consistent Dynamic Stereo Matching
Junpeng Jing, Ye Mao, and Krystian Mikolajczyk
DisC-Diff: Disentangled Conditional Diffusion Model for Multi-Contrast MRI Super-Resolution
Ye Mao, Lan Jiang, Xi Chen, and Chao Li
CoLa-Diff: Conditional Latent Diffusion Model for Multi-Modal MRI Synthesis
Lan Jiang, Ye Mao, Xi Chen, and Chao Li
Deep Domain Adaptation Enhances Amplification Curve Analysis for Single-Channel Multiplexing in Real-Time PCR
Ye Mao, Ke Xu, Luca Miglietta, Louis Kreitmann, Nicolas Moser, Pantelis Georgiou, Alison Holmes, and Jesus Rodriguez-Manzano
Academic Record
Experience
Education, research appointments, teaching, service, and technical strengths.
Education
- PhD (Probationary) in Computer Vision and Machine Learning Imperial College London, 2023-present
- MPhil in Medical Sciences University of Cambridge, 2022-2023 Thesis: Brain MRI Super-Resolution using Conditional Diffusion Model
- MSc in Applied Machine Learning Imperial College London, 2021-2022 First Class Honours, top overall grade: 81%; thesis grade: 87% Thesis: Domain Adaptation for Digital PCR Multiplexing
- BSc in Computer Science King's College London, 2018-2021 First Class Honours, top overall grade: 82%; thesis grade: 81% Thesis: String Sanitisation Algorithm Development
Research Experience
- Research Assistant Centre for Antimicrobial Optimisation Lab Imperial College London, 2024-2025
- Research Assistant Cambridge Brain Tumour Imaging Lab University of Cambridge, January 2023-September 2023
Teaching
-
Graduate Teaching Assistant
Imperial College London, 2024-present
- Deep Learning (ELEC60009)
- Computer Vision and Pattern Recognition (ELEC70073)
-
Undergraduate Teaching Assistant
King's College London, 2021
- Foundation of Computing
- Robotics Group Project
Academic Service
- Computer Vision CVPR (2025-2026), ECCV (2024, 2026), ICCV (2025)
- Machine Learning NeurIPS (2024-2026), ICML (2025-2026), ICLR (2025-2026)
- Medical Imaging IEEE Transactions on Medical Imaging, MICCAI (2024-2026)
Professional Skills
- Tools Python, PyTorch, TensorFlow, C++, Java, OpenCV, Git, LaTeX
- Languages Mandarin (native), English (fluent)
Recognition
Awards
Scholarships and prizes recognising academic excellence across research and study.
- Imperial President’s PhD Scholarship Among the top 50 selected college-wide, 2023
- The Humanitarian Trust GBP 1,000 funding to support MPhil study in Clinical Neuroscience
- Applied Machine Learning Prize Highest overall grade in MSc Applied Machine Learning, 2022
- Hertha Ayrton Centenary Prize Highest final-year project grade in the Imperial EEE department, 2022
- Robotics Prize Highest overall grade in BSc Computer Science and Robotics, 2021
- Peplow Prize Highest final-year project grade in King's Informatics department, 2021