Current
Reinforcement Learning for Traffic Signal Control
Deep and multi-agent reinforcement learning for intelligent traffic signal control, trained in rich 3D simulation and transferred to real-world networks.
George Vogiatzis
Reader in Computer Science, Loughborough University
I work in computer vision and deep learning — including 3D reconstruction, self-supervised monocular depth estimation, reinforcement learning for intelligent transport, and vision and language. Browse my research under Projects, find my papers under Publications, read more About my background, or get in touch via Contact.
Research interests
3D ReconstructionRecovering shape and structure from images.Computer VisionImage and video understanding, geometry, and recognition.Intelligent Transport SystemsVision and learning for traffic.Machine LearningDeep learning methods for vision and beyond.Monocular Depth EstimationSelf-supervised depth from single images.Photometric StereoSurface normals and 3D shape from shading.Reinforcement LearningSequential decision making.
Featured projects
Completed
Self-Supervised Monocular Depth Estimation
Learning to predict depth from single images without ground-truth supervision, with a focus on dynamic scenes and challenging conditions.
Completed
3D Reconstruction and Multi-View Stereo
Recovering accurate 3D shape from collections of images, using multi-view stereo, volumetric graph-cuts, and probabilistic depth-map fusion.
Recent publications
- PUVC: A novel position-based VANET caching framework built upon UAVs associationLoughborough University Research Repository (Loughborough University) · 2026
- When Memory Matters: An Evaluation of LSTM-Based Multi-Agent Learning for Multi-Intersection Traffic Signal ControlWhite Rose Research Online (University of Leeds, The University of Sheffield, University of York) · 2026
- Multi-agent LLMs for Closed Loop Supply Chain implementation (Extended Abstract)Proceedings of the AAAI Symposium Series · 2026
- PSNVC: A Novel Physical-Based Vanet Caching Framework by Integrating SDN and NDN Paradigm2025
- Ssvc: A Novel Vanet Caching Framework of Integrating Network Slicing and Spiking Neural NetworkSSRN Electronic Journal · 2025