Robust Segmentation

The delineation of regions in medical images - segmentation - is a key task across modalities and organs. Methods developed need to produce segmentations that are biologically meaningful across scanners and imaging protocols.

Home / Harmonisation / Segmentation / Translation / Privacy / Explainable AI

Papers

	Automated quality assessment using appearance-based simulations and hippocampus segmentation on Low-field paediatric brain MR images Vaanathi Sundaresan, Nicola K Dinsdale Low Field Pediatric Brain Magnetic Resonance Image Segmentation and Quality Assurance Challenge, 2024 Paper / Code Winner of LISA Challenge 2024 @ MICCAI 2024
	Is Your Style Transfer Doing Anything Useful? An Investigation Into Hippocampus Segmentation and the Role of Preprocessing Hoda Kalabizadeh, Ludovica Griffanti, Pak-Hei Yeung, Natalie Voets, Grace Gillis, Clare E Mackay, Ana IL Namburete, Nicola K Dinsdale, Konstantinos Kamnitsas Machine Learning in Clinical Neuroimaging, 2024 Paper We investigated the performance of segmentation models trained on research data that were style-transferred to resemble clinical scans. Our results highlighted the importance of intensity normalisation methods in MRI segmentation, and their relation to domain shift and style-transfer.
	Anatomically plausible segmentations: Explicitly preserving topology through prior deformations Madeleine K Wyburd, Nicola K Dinsdale , Ana IL Namburete, Mark Jenkinson Medical Image Analysis 2024 Paper / Code Our model, TEDS-Net, generates anatomically plausible segmentations through deforming a prior shape with the same topology as the anatomy of interest.
	UniFed: A unified deep learning framework for segmentation of partially labelled, distributed neuroimaging data Nicola K Dinsdale, Mark Jenkinson, Ana IL Namburete bioRxiv, 2024 Project Page / Paper / Code We, therefore, propose UniFed, a unified federated harmoni- sation framework, which enables three key processes to be completed: 1) the training of a federated partially labelled harmonisation network, 2) the selection of the most appropriate pretrained model for a new unseen site, and 3) the incorporation of a new site into the harmonised federation.
	STAMP: Simultaneous Training and Model Pruning for Low Data Regimes in Medical Image Segmentation Nicola K Dinsdale , Mark Jenkinson, Ana IL Namburete Medical Image Analysis, 2022 Project Page / Paper / Code Development of an algorithm to enable simultaneous training and pruning, enabling segmentation in low data domains. Also introduces Targeted Dropout to stabilise the pruning.
	Omni-Supervised Domain Adversarial Training for White Matter Hyperintensity Segmentation in the UK Biobank Vaanathi Sundaresan, Nicola K Dinsdale, Ludovica Griffanti, Mark Jenkinson, ISBI 2022 [Oral Presentation] Project Page / Paper / Code Exploring the use of omni-supervised learning for white matter hyperintensity segmentation, using age prediction as the selection criteria. Leads to significant increase in sensitivity especially for small lesions.
	The impact of transfer learning on 3D deep learning convolutional neural network segmentation of the hippocampus in mild cognitive impairment and Alzheimer disease subjects Erica Balboni, Luca Nocetti, Chiara Carbone, Nicola K Dinsdale , Maurilio Genovese, Gabriele Guidi, Marcella Malagoli, Annalisa Chiari, Ana IL Namburete, Mark Jenkinson, Giovanna Zamboni, Human Brain Mapping 2022 Paper / Code Exploration of the use of my SWANS algorithm for clinical data, across disease groups - SWANS first presented at MICCAI 2019
	TEDS-Net: Enforcing Diffeomorphisms in Spatial Transformers to Guarantee Topology Preservation in Segmentations Madeleine K Wyburd, Nicola K Dinsdale , Ana IL Namburete, Mark Jenkinson MICCAI, 2021 [Early Acceptance] Project Page / Paper / Code Novel segmentation method guaranteeing accurate topology.
	Comparison of domain adaptation techniques for white matter hyperintensity segmentation in brain MR images Vaanathi Sundaresan, Giovanna Zamboni, Nicola K Dinsdale , Peter Rothwell, Ludovica Griffanti, Mark Jenkinson Medical Image Analysis 2021 Paper / Code Comparison of DA methods for white matter hyperintensity segmentation, comparing methods including my proposed method: paper.
	A 2-step deep learning method with domain adaptation for multi-centre, multi-vendor and multi-disease cardiac magnetic resonance segmentation Jorge Corral Acero, Vaanathi Sundaresan, Nicola K Dinsdale , Vicente Grau, Mark Jenkinson STACOM 2020 - 6th place Paper Segmentation of cardiac MRI across sites, using the method proposed in my paper
	Spatial warping network for 3D segmentation of the hippocampus in MR images Nicola K Dinsdale , Mark Jenkinson, Ana IL Namburete MICCAI, 2019 [Early Acceptance] Paper / Code Novel segmentation method based on a spatial transformer network.

The template of this webpage is from source code.