collaborators

6 papers

cs.CV2026

Label tree semantic losses for rich multi-class medical image segmentation

Junwen Wang, Oscar MacCormac, William Rochford +3

Rich and accurate medical image segmentation is poised to underpin the next generation of AI-defined clinical practice by delineating critical anatomy for pre-operative planning, g…

cs.CV2025

Longitudinal Vestibular Schwannoma Dataset with Consensus-based Human-in-the-loop Annotations

Navodini Wijethilake, Marina Ivory, Oscar MacCormac +17

Accurate segmentation of vestibular schwannoma (VS) on Magnetic Resonance Imaging (MRI) is essential for patient management but often requires time-intensive manual annotations by…

cs.CV2025

Beyond one-hot encoding? Journey into compact encoding for large multi-class segmentation

Aaron Kujawa, Thomas Booth, Tom Vercauteren

This work presents novel methods to reduce computational and memory requirements for medical image segmentation with a large number of classes. We curiously observe challenges in m…

eess.IV2025

crossMoDA Challenge: Evolution of Cross-Modality Domain Adaptation Techniques for Vestibular Schwannoma and Cochlea Segmentation from 2021 to 2023

Navodini Wijethilake, Reuben Dorent, Marina Ivory +38

The cross-Modality Domain Adaptation (crossMoDA) challenge series, initiated in 2021 in conjunction with the International Conference on Medical Image Computing and Computer Assist…

cs.CV2025

Tree-based Semantic Losses: Application to Sparsely-supervised Large Multi-class Hyperspectral Segmentation

Junwen Wang, Oscar Maccormac, William Rochford +3

Hyperspectral imaging (HSI) shows great promise for surgical applications, offering detailed insights into biological tissue differences beyond what the naked eye can perceive. Ref…

eess.IV2025

A generalisable head MRI defacing pipeline: Evaluation on 2,566 meningioma scans

Lorena Garcia-Foncillas Macias, Aaron Kujawa, Aya Elshalakany +2

Reliable MRI defacing techniques to safeguard patient privacy while preserving brain anatomy are critical for research collaboration. Existing methods often struggle with incomplet…