Develops computer vision methods for 3D reconstruction, visual SLAM and image analysis using convex optimization, statistical learning and neural networks.
3D vision & neural rendering
Learning theory & optimization
2026BA-T: An Iterative Transformer for Two-View Bundle AdjustmentNeurIPS
2026Combining Projected Uncertainty for Self-Supervised Visual Odometry: From Two-Frame to Multi-FrameIJCV
75 T1 · 138 papers · 35 members · 45 alumni
GBOxfordDepartment of Engineering Science
Andrea VedaldiProfessor of Computer Vision and Machine Learning
Develops computer vision methods that understand image and video semantics and 3D geometry with little or no manual supervision.
3D vision & neural rendering
Video understanding
2026CoWTracker: Tracking by Warping instead of CorrelationCVPR
2026Free-Range Gaussians: Non-Grid-Aligned Generative 3D Gaussian ReconstructionECCV
74 T1 · 86 papers · Group size not listed
CNZhejiangCollege of Computer Science and Technology
Researches 3D digitization at the intersection of graphics, vision and AI, including 3D reconstruction from RGB-D video and photorealistic digital humans.
3D vision & neural rendering
Video understanding
Generative 3D & visual content
2026Animating the Uncaptured: Humanoid Mesh Animation with Video Diffusion ModelsICLR
2026Diff3R: Feed-forward 3D Gaussian Splatting with Uncertainty-aware Differentiable OptimizationNeurIPS
72 T1 · 77 papers · 18 members
CHEPFLSchool of Computer and Communication Sciences (IC)
Leads the Computer Vision Laboratory, working on 3D reconstruction, shape modelling, biomedical image analysis and computer-assisted engineering design.
3D vision & neural rendering
Medical image analysis
Signal processing & information theory
2026DiffAtlas: GenAI-Fying Atlas Segmentation via Image-Mask DiffusionMICCAI
2026End-to-End 4D Heart Mesh Recovery Across Full-Stack and Sparse Cardiac MRITMLR
Builds vision foundation models, including repurposing generative models for perception tasks (DiCeption, GenPercept) and metric depth estimation (Metric3D).
3D vision & neural rendering
Generative & foundation models
Robot perception & SLAM
2026ACTIVE-o3: Empowering MLLMs with Active Perception via Pure Reinforcement LearningICML
2026Efficient Self-Evaluation for Diffusion Language Models via Sequence RegenerationACL
Works on 3D scene understanding and reconstruction from commodity RGB-D data, using generative 3D deep learning to create semantically grounded 3D environments.
3D vision & neural rendering
Recognition, detection & segmentation
Generative 3D & visual content
2026Animating the Uncaptured: Humanoid Mesh Animation with Video Diffusion ModelsICLR
2026Diff3R: Feed-forward 3D Gaussian Splatting with Uncertainty-aware Differentiable OptimizationNeurIPS