Deep Learning Methodology

Fundemental to to our research is the developement of state of the art DL methodology. Key topics include segmentation, interpretability and domain adpatation.

Segmentation

Anatomically plausible segmentations: Explicitly preserving topology through prior deformations

Wyburd et al. Medical Image Analysis (2024)

Sli2Vol: Annotate a 3D Volume from a Single Slice with Self-supervised Learning

Yeung et al. MICCAI (2021)

STAMP: Simultaneous Training and Model Pruning for Low Data Regimes in Medical Image Segmentation

Dinsdale et al. Medical Image Analysis (2022)

Uncertainty Estimates as Data Selection Criteria to Boost Omni-Supervised Learning

Venturini et al. MICCAI (2020)

Spatial Warping Network for 3D Segmentation of the Hippocampus in MR Images

Dinsdale et al. MICCAI (2019)

Improving U-Net Segmentation with Active Contour Based Label Correction

Hesse et al. MIUA (2020)

TEDS-Net: Enforcing Diffeomorphisms in Spatial Transformers to Guarantee Topology Preservation in Segmentations

Wyburd et al. MICCAI (2021)

Domain Adaptation

SFHarmony: Source Free Domain Adaptation for Distributed Neuroimaging Analysis

Dinsdale et al. ICCV (2023)

FedHarmony: Unlearning Scanner Bias with Distributed Data

Dinsdale et al. MICCAI (2022)

Deep learning-based unlearning of dataset bias for MRI harmonisation and confound removal

Dinsdale et al. NeuroImage (2021)

Intepretability

Prototype learning for explainable brain age prediction

Hesse et al. WACV (2024)

INSightR-Net: Interpretable Neural Network for Regression Using Similarity-Based Comparisons to Prototypical Examples

Hesse et al. MICCAI (2023)