This paper proposes a federated learning model for COVID-19 diagnosis using CT scans. Data remains decentralized across institutions, preserving patient privacy. The architecture ensures accurate detection by collaboratively training models across diverse datasets without compromising sensitive medical imaging information.

Fiona Isabelle Chamberlain, Julian Eric Thorne, Lauren Michelle Prescott, Dominic Wesley Shaw, Nathan Charles Liddell

This research integrates neuro-symbolic AI with case-based reasoning to enhance diagnosis of rare genetic disorders. Using multimodal clinical images and structured knowledge, the framework improves precision by combining deep visual recognition with logical inference for complex phenotype-based disease classification.

Sophie Eleanor Marsh, Henry Jacob Lindholm, Evelyn Theresa Cowell, Christopher David Holtz, Isabelle Frances Chapman

This study presents a lightweight AI framework deployed across edge and cloud for continuous monitoring of COPD patients. By analyzing physiological sensor data in real-time, it supports early anomaly detection, patient alerts, and personalized intervention, reducing hospital readmissions and mortality rates.

Matilda Grace Everly, Patrick Lawrence Vaughn, Oliver Dean Whitaker, Georgia Faith Kenyon, Frederick Ross Maynard

GANs are used to generate synthetic retinal images to augment training datasets for diabetic retinopathy classification. The enhanced data diversity improves model generalization, reducing overfitting and improving diagnostic performance in deep learning models applied to limited ophthalmic datasets.

Florence Margaret Reeves, Sebastian Hugo Tate, Clara Joanne Huntington, Maxwell Finn Andrews, Emily Sophia Doran

This work introduces a cloud-native NLP pipeline for summarizing multilingual radiology reports. The framework leverages transformer-based models to extract key findings, enabling efficient decision support and reducing cognitive load for clinicians operating in high-volume diagnostic environments across global clinical data warehouses.

Edward Nathaniel Bromley, Louise Catherine Sanderson, Charlotte Ivy Mulholland, Alexander George Tripp, Lydia Beatrice Redmond

This paper presents a deep learning model for histopathological image segmentation, focusing on breast cancer tissue analysis. By detecting tumor regions accurately, the system aids pathologists in diagnosis and treatment planning, enhancing reproducibility and reducing diagnostic variability in clinical practice.

Jeremy Thomas Calderon, Rebecca Anne Vickers, Gabriel Vincent Morrison, Olivia Madeleine Blake, Dominic James Holcombe

This study introduces a context-aware system combining semantic sensor fusion and edge computing for health monitoring in smart homes. The framework enables real-time detection of abnormal activities and physiological anomalies, offering personalized and privacy-preserving care for elderly and chronic patients.

Isla Harriet Donnelly, Felix George Mowbray, Eleanor Lucy Dalrymple, Benjamin Oscar Fielding, Harriet Sophie Langston