Research Students

Name: Nicholas Geere

PhD/project title: Monitoring and Prediction of Rail Defects using Multimodal Neural Networks

Research description: Identifying and predicting the development of squats, corrugation and RCF faults in high-speed rail using multimodal Neural Networks. The modalities considered include digital images, ultrasound scans, track geometry data, maintenance records and tonnage.

Name: Paul Richards

PhD/project title: Using Machine Learning to understand the role of the Soil Microbiome in Carbon Sequestration

Research description: This interdisciplinary project develops a machine learning model to connect diverse aspects of soil—its structure, microbial DNA, organismal diversity, chemistry, and enzymatic activity. The model enables predictions from limited data, advancing climate change mitigation and deepening our understanding of soil ecosystem interactions.

Name: James Wright

PhD/project title: Using Generative AI to improve segmentation effectiveness and quality of imaging-derived features in CT and MRI images

Research description: My research uses diffusion models to solve two important problems in medical AI: the lack of expert-annotated data and generalisability issues caused by variations between scanners. By creating high-quality synthetic data, we are aiming to improve the adaptability and generalisability of AI models for medical image segmentation.

Name: Zihe (Zach) Gao

PhD/project title: 3D Model Generation for Plants

Research description: Generating 3D models or 3D meshes reconstruction from single images is an intriguing, important, but difficult problem, especially for plant research. This research will be helpful for training image-based AI models, for things like disease identification and other phenotyping tasks, therefore enhancing agricultural and ecological studies.

Name: Lewis Stuart

PhD Title: Intuitive Reconstruction and Synthetic Generation of High-Fidelity 3D Wheat Plant Models

Research description: Representing wheat plants in 3D remains a bottleneck in high-throughput phenotyping pipelines. This project aims to investigate state-of-the-art reconstruction methods for capturing accurate 3D representations of wheat, while also exploring approaches to synthetically generate realistic 3D wheat plant models that faithfully replicate the morphology and texturing of real plants.

Name: Amber Swarbrick

PhD Title: Environment-Aware Optimisation for High-Resolution 3D Reconstruction of Plant Roots

Research description: Despite advances in 3D root reconstruction, state-of-the-art methods are still hindered by high computational overhead. This project aims to integrate environmental and physiological modelling into root reconstruction frameworks. By leveraging plant growth predictions and environmental data, the project aims to enhance computational efficiency, enabling faster, more precise root system analysis for improved crop resilience in the face of climate change.