James Willoughby
James Willoughby
Wolfson Building, Parks Road, Oxford OX1 3QD
Interests
I am a DPhil in Computer Science, working on the use of Deep Learning in Medical Image Analysis. Most of my work focuses on Segmentation of medically relevant features across a wide range of modalities and tissues, but I have previously worked on Object Detection and Image Generation.
One of my main interests is in uncertainty-based methods for medical image analysis. Owing to the high level of responsibility required in making clinical decisions, methods which incorporate and communicate uncertainty effectively are critical to creating pipelines which both clinicians and patients feel safe using.
I have also previously worked as a research assistant at the UCL Institute of Ophthalmology looking at the use of Deep Learning in Classification and Localisation of Disease Markers in OCT scans.
Selected Publications
-
Runtime Freezing: Dynamic Class Loss for multi−organ 3D Segmentation
James Willoughby and Irina Voiculescu
In IEEE International Symposium on Biomedical Imaging (ISBI). May, 2024.
Details about Runtime Freezing: Dynamic Class Loss for multi−organ 3D Segmentation | BibTeX data for Runtime Freezing: Dynamic Class Loss for multi−organ 3D Segmentation | Link to Runtime Freezing: Dynamic Class Loss for multi−organ 3D Segmentation