-
SFHarmony: Source Free Domain Adaptation for Distributed Neuroimaging Analysis
Authors:
Nicola K Dinsdale,
Mark Jenkinson,
Ana IL Namburete
Abstract:
To represent the biological variability of clinical neuroimaging populations, it is vital to be able to combine data across scanners and studies. However, different MRI scanners produce images with different characteristics, resulting in a domain shift known as the `harmonisation problem'. Additionally, neuroimaging data is inherently personal in nature, leading to data privacy concerns when shari…
▽ More
To represent the biological variability of clinical neuroimaging populations, it is vital to be able to combine data across scanners and studies. However, different MRI scanners produce images with different characteristics, resulting in a domain shift known as the `harmonisation problem'. Additionally, neuroimaging data is inherently personal in nature, leading to data privacy concerns when sharing the data. To overcome these barriers, we propose an Unsupervised Source-Free Domain Adaptation (SFDA) method, SFHarmony. Through modelling the imaging features as a Gaussian Mixture Model and minimising an adapted Bhattacharyya distance between the source and target features, we can create a model that performs well for the target data whilst having a shared feature representation across the data domains, without needing access to the source data for adaptation or target labels. We demonstrate the performance of our method on simulated and real domain shifts, showing that the approach is applicable to classification, segmentation and regression tasks, requiring no changes to the algorithm. Our method outperforms existing SFDA approaches across a range of realistic data scenarios, demonstrating the potential utility of our approach for MRI harmonisation and general SFDA problems. Our code is available at \url{https://github.com/nkdinsdale/SFHarmony}.
△ Less
Submitted 28 March, 2023;
originally announced March 2023.
-
Segmentation method for cerebral blood vessels from MRA using hysteresis
Authors:
Georgia Kenyon,
Stephan Lau,
Michael A. Chappell,
Mark Jenkinson
Abstract:
Segmentation of cerebral blood vessels from Magnetic Resonance Imaging (MRI) is an open problem that could be solved with deep learning (DL). However, annotated data for training is often scarce. Due to the absence of open-source tools, we aim to develop a classical segmentation method that generates vessel ground truth from Magnetic Resonance Angiography for DL training of segmentation across a v…
▽ More
Segmentation of cerebral blood vessels from Magnetic Resonance Imaging (MRI) is an open problem that could be solved with deep learning (DL). However, annotated data for training is often scarce. Due to the absence of open-source tools, we aim to develop a classical segmentation method that generates vessel ground truth from Magnetic Resonance Angiography for DL training of segmentation across a variety of modalities. The method combines size-specific Hessian filters, hysteresis thresholding and connected component correction. The optimal choice of processing steps was evaluated with a blinded scoring by a clinician using 24 3D images. The results show that all method steps are necessary to produce the highest (14.2/15) vessel segmentation quality score. Omitting the connected component correction caused the largest quality loss. The method, which is available on GitHub, can be used to train DL models for vessel segmentation.
△ Less
Submitted 9 March, 2023;
originally announced March 2023.
-
FedHarmony: Unlearning Scanner Bias with Distributed Data
Authors:
Nicola K Dinsdale,
Mark Jenkinson,
Ana IL Namburete
Abstract:
The ability to combine data across scanners and studies is vital for neuroimaging, to increase both statistical power and the representation of biological variability. However, combining datasets across sites leads to two challenges: first, an increase in undesirable non-biological variance due to scanner and acquisition differences - the harmonisation problem - and second, data privacy concerns d…
▽ More
The ability to combine data across scanners and studies is vital for neuroimaging, to increase both statistical power and the representation of biological variability. However, combining datasets across sites leads to two challenges: first, an increase in undesirable non-biological variance due to scanner and acquisition differences - the harmonisation problem - and second, data privacy concerns due to the inherently personal nature of medical imaging data, meaning that sharing them across sites may risk violation of privacy laws. To overcome these restrictions, we propose FedHarmony: a harmonisation framework operating in the federated learning paradigm. We show that to remove the scanner-specific effects, we only need to share the mean and standard deviation of the learned features, hel** to protect individual subjects' privacy. We demonstrate our approach across a range of realistic data scenarios, using real multi-site data from the ABIDE dataset, thus showing the potential utility of our method for MRI harmonisation across studies. Our code is available at https://github.com/nkdinsdale/FedHarmony.
△ Less
Submitted 31 May, 2022;
originally announced May 2022.
-
Uncertainty categories in medical image segmentation: a study of source-related diversity
Authors:
Luke Whitbread,
Mark Jenkinson
Abstract:
Measuring uncertainties in the output of a deep learning method is useful in several ways, such as in assisting with interpretation of the outputs, hel** build confidence with end users, and for improving the training and performance of the networks. Several different methods have been proposed to estimate uncertainties, including those from epistemic (relating to the model used) and aleatoric (…
▽ More
Measuring uncertainties in the output of a deep learning method is useful in several ways, such as in assisting with interpretation of the outputs, hel** build confidence with end users, and for improving the training and performance of the networks. Several different methods have been proposed to estimate uncertainties, including those from epistemic (relating to the model used) and aleatoric (relating to the data) sources using test-time dropout and augmentation, respectively. Not only are these uncertainty sources different, but they are governed by parameter settings (e.g., dropout rate or type and level of augmentation) that establish even more distinct uncertainty categories. This work investigates how different the uncertainties are from these categories, for magnitude and spatial pattern, to empirically address the question of whether they provide usefully distinct information that should be captured whenever uncertainties are used. We take the well characterised BraTS challenge dataset to demonstrate that there are substantial differences in both magnitude and spatial pattern of uncertainties from the different categories, and discuss the implications of these in various use cases.
△ Less
Submitted 16 September, 2022; v1 submitted 1 March, 2022;
originally announced March 2022.
-
Mutual information neural estimation for unsupervised multi-modal registration of brain images
Authors:
Gerard Snaauw,
Michele Sasdelli,
Gabriel Maicas,
Stephan Lau,
Johan Verjans,
Mark Jenkinson,
Gustavo Carneiro
Abstract:
Many applications in image-guided surgery and therapy require fast and reliable non-linear, multi-modal image registration. Recently proposed unsupervised deep learning-based registration methods have demonstrated superior performance compared to iterative methods in just a fraction of the time. Most of the learning-based methods have focused on mono-modal image registration. The extension to mult…
▽ More
Many applications in image-guided surgery and therapy require fast and reliable non-linear, multi-modal image registration. Recently proposed unsupervised deep learning-based registration methods have demonstrated superior performance compared to iterative methods in just a fraction of the time. Most of the learning-based methods have focused on mono-modal image registration. The extension to multi-modal registration depends on the use of an appropriate similarity function, such as the mutual information (MI). We propose guiding the training of a deep learning-based registration method with MI estimation between an image-pair in an end-to-end trainable network. Our results show that a small, 2-layer network produces competitive results in both mono- and multi-modal registration, with sub-second run-times. Comparisons to both iterative and deep learning-based methods show that our MI-based method produces topologically and qualitatively superior results with an extremely low rate of non-diffeomorphic transformations. Real-time clinical application will benefit from a better visual matching of anatomical structures and less registration failures/outliers.
△ Less
Submitted 6 October, 2022; v1 submitted 25 January, 2022;
originally announced January 2022.
-
TEDS-Net: Enforcing Diffeomorphisms in Spatial Transformers to Guarantee Topology Preservation in Segmentations
Authors:
Madeleine K. Wyburd,
Nicola K. Dinsdale,
Ana I. L. Namburete,
Mark Jenkinson
Abstract:
Accurate topology is key when performing meaningful anatomical segmentations, however, it is often overlooked in traditional deep learning methods. In this work we propose TEDS-Net: a novel segmentation method that guarantees accurate topology. Our method is built upon a continuous diffeomorphic framework, which enforces topology preservation. However, in practice, diffeomorphic fields are represe…
▽ More
Accurate topology is key when performing meaningful anatomical segmentations, however, it is often overlooked in traditional deep learning methods. In this work we propose TEDS-Net: a novel segmentation method that guarantees accurate topology. Our method is built upon a continuous diffeomorphic framework, which enforces topology preservation. However, in practice, diffeomorphic fields are represented using a finite number of parameters and sampled using methods such as linear interpolation, violating the theoretical guarantees. We therefore introduce additional modifications to more strictly enforce it. Our network learns how to warp a binary prior, with the desired topological characteristics, to complete the segmentation task. We tested our method on myocardium segmentation from an open-source 2D heart dataset. TEDS-Net preserved topology in 100% of the cases, compared to 90% from the U-Net, without sacrificing on Hausdorff Distance or Dice performance. Code will be made available at: www.github.com/mwyburd/TEDS-Net
△ Less
Submitted 28 July, 2021;
originally announced July 2021.
-
Challenges for machine learning in clinical translation of big data imaging studies
Authors:
Nicola K Dinsdale,
Emma Bluemke,
Vaanathi Sundaresan,
Mark Jenkinson,
Stephen Smith,
Ana IL Namburete
Abstract:
The combination of deep learning image analysis methods and large-scale imaging datasets offers many opportunities to imaging neuroscience and epidemiology. However, despite the success of deep learning when applied to many neuroimaging tasks, there remain barriers to the clinical translation of large-scale datasets and processing tools. Here, we explore the main challenges and the approaches that…
▽ More
The combination of deep learning image analysis methods and large-scale imaging datasets offers many opportunities to imaging neuroscience and epidemiology. However, despite the success of deep learning when applied to many neuroimaging tasks, there remain barriers to the clinical translation of large-scale datasets and processing tools. Here, we explore the main challenges and the approaches that have been explored to overcome them. We focus on issues relating to data availability, interpretability, evaluation and logistical challenges, and discuss the challenges we believe are still to be overcome to enable the full success of big data deep learning approaches to be experienced outside of the research field.
△ Less
Submitted 7 July, 2021;
originally announced July 2021.
-
Self-supervised Lesion Change Detection and Localisation in Longitudinal Multiple Sclerosis Brain Imaging
Authors:
Minh-Son To,
Ian G Sarno,
Chee Chong,
Mark Jenkinson,
Gustavo Carneiro
Abstract:
Longitudinal imaging forms an essential component in the management and follow-up of many medical conditions. The presence of lesion changes on serial imaging can have significant impact on clinical decision making, highlighting the important role for automated change detection. Lesion changes can represent anomalies in serial imaging, which implies a limited availability of annotations and a wide…
▽ More
Longitudinal imaging forms an essential component in the management and follow-up of many medical conditions. The presence of lesion changes on serial imaging can have significant impact on clinical decision making, highlighting the important role for automated change detection. Lesion changes can represent anomalies in serial imaging, which implies a limited availability of annotations and a wide variety of possible changes that need to be considered. Hence, we introduce a new unsupervised anomaly detection and localisation method trained exclusively with serial images that do not contain any lesion changes. Our training automatically synthesises lesion changes in serial images, introducing detection and localisation pseudo-labels that are used to self-supervise the training of our model. Given the rarity of these lesion changes in the synthesised images, we train the model with the imbalance robust focal Tversky loss. When compared to supervised models trained on different datasets, our method shows competitive performance in the detection and localisation of new demyelinating lesions on longitudinal magnetic resonance imaging in multiple sclerosis patients. Code for the models will be made available on GitHub.
△ Less
Submitted 1 June, 2021;
originally announced June 2021.
-
Brain tumour segmentation using a triplanar ensemble of U-Nets
Authors:
Vaanathi Sundaresan,
Ludovica Griffanti,
Mark Jenkinson
Abstract:
Gliomas appear with wide variation in their characteristics both in terms of their appearance and location on brain MR images, which makes robust tumour segmentation highly challenging, and leads to high inter-rater variability even in manual segmentations. In this work, we propose a triplanar ensemble network, with an independent tumour core prediction module, for accurate segmentation of these t…
▽ More
Gliomas appear with wide variation in their characteristics both in terms of their appearance and location on brain MR images, which makes robust tumour segmentation highly challenging, and leads to high inter-rater variability even in manual segmentations. In this work, we propose a triplanar ensemble network, with an independent tumour core prediction module, for accurate segmentation of these tumours and their sub-regions. On evaluating our method on the MICCAI Brain Tumor Segmentation (BraTS) challenge validation dataset, for tumour sub-regions, we achieved a Dice similarity coefficient of 0.77 for both enhancing tumour (ET) and tumour core (TC). In the case of the whole tumour (WT) region, we achieved a Dice value of 0.89, which is on par with the top-ranking methods from BraTS'17-19. Our method achieved an evaluation score that was the equal 5th highest value (with our method ranking in 10th place) in the BraTS'20 challenge, with mean Dice values of 0.81, 0.89 and 0.84 on ET, WT and TC regions respectively on the BraTS'20 unseen test dataset.
△ Less
Submitted 24 May, 2021;
originally announced May 2021.
-
High-Order Accurate FDTD Schemes for Dispersive Maxwell's Equations in Second-Order Form Using Recursive Convolutions
Authors:
Michael J. Jenkinson,
Jeffrey W. Banks
Abstract:
We propose a novel finite-difference time-domain (FDTD) scheme for the solution of the Maxwell's equations in which linear dispersive effects are present. The method uses high-order accurate approximations in space and time for the dispersive Maxwell's equations written as a second-order vector wave equation with a time-history convolution term. The modified equation approach is combined with the…
▽ More
We propose a novel finite-difference time-domain (FDTD) scheme for the solution of the Maxwell's equations in which linear dispersive effects are present. The method uses high-order accurate approximations in space and time for the dispersive Maxwell's equations written as a second-order vector wave equation with a time-history convolution term. The modified equation approach is combined with the recursive convolution (RC) method to develop high-order approximations accurate to any desired order in space and time. High-order-accurate centered approximations of the physical Maxwell interface conditions are derived for the dispersive setting in order to fully restore accuracy at discontinuous material interfaces. Second- and fourth-order accurate versions of the scheme are presented and implemented in two spatial dimensions for the case of the Drude linear dispersion model. The stability of these schemes is analyzed. Finally, our approach is also amenable to curvilinear numerical grids if used with appropriate generalized Laplace operator.
△ Less
Submitted 14 June, 2017;
originally announced June 2017.
-
Discrete Solitary Waves in Systems with Nonlocal Interactions and the Peierls-Nabarro Barrier
Authors:
Michael Jenkinson,
Michael I. Weinstein
Abstract:
We study a class of discrete focusing nonlinear Schr{ö}dinger equations (DNLS) with general nonlocal interactions. We prove the existence of onsite and offsite discrete solitary waves, which bifurcate from the trivial solution at the endpoint frequency of the continuous spectrum of linear dispersive waves. We also prove exponential smallness, in the frequency-distance to the bifurcation point, of…
▽ More
We study a class of discrete focusing nonlinear Schr{ö}dinger equations (DNLS) with general nonlocal interactions. We prove the existence of onsite and offsite discrete solitary waves, which bifurcate from the trivial solution at the endpoint frequency of the continuous spectrum of linear dispersive waves. We also prove exponential smallness, in the frequency-distance to the bifurcation point, of the Peierls-Nabarro energy barrier (PNB), as measured by the difference in Hamiltonian or mass functionals evaluated on the onsite and offsite states. These results extend those of the authors for the case of nearest neighbor interactions to a large class of nonlocal short-range and long-range interactions. The appearance of distinct onsite and offsite states is a consequence of the breaking of continuous spatial translation invariance. The PNB plays a role in the dynamics of energy transport in such nonlinear Hamiltonian lattice systems.
Our class of nonlocal interactions is defined in terms of coupling coefficients, $J_m$, where $m\in\mathbb{Z}$ is the lattice site index, with $J_m\simeq m^{-1-2s}, s\in[1,\infty)$ and $J_m\sim e^{-γ|m|},\ s=\infty,\ γ>0,$ (Kac-Baker). For $s\ge1$, the bifurcation is seeded by solutions of the (effective / homogenized) cubic focusing nonlinear Schr{ö}dinger equation (NLS). However, for $1/4<s<1$, the bifurcation is controlled by the fractional nonlinear Schr{ö}dinger equation, FNLS, with $(-Δ)^s$ replacing $-Δ$. The proof is based on a Lyapunov-Schmidt reduction strategy applied to a momentum space formulation. The PN barrier bounds require appropriate uniform decay estimates for the discrete Fourier transform of DNLS discrete solitary waves. A key role is also played by non-degeneracy of the ground state of FNLS, recently proved by Frank, Lenzmann \& Silvestre.
△ Less
Submitted 9 January, 2017; v1 submitted 18 January, 2016;
originally announced January 2016.
-
On-Site and Off-Site Bound States of the Discrete Nonlinear Schrödinger Equation and the Peierls-Nabarro Barrier
Authors:
Michael Jenkinson,
Michael I. Weinstein
Abstract:
We construct multiple families of solitary standing waves of the discrete cubically nonlinear Schrödinger equation (DNLS) in dimensions $d=1,2$ and $3$. These states are obtained via a bifurcation analysis about the continuum (NLS) limit. One family consists {\it on-site symmetric} (vertex-centered) states; these are spatially localized solitary standing waves which are symmetric about any fixed l…
▽ More
We construct multiple families of solitary standing waves of the discrete cubically nonlinear Schrödinger equation (DNLS) in dimensions $d=1,2$ and $3$. These states are obtained via a bifurcation analysis about the continuum (NLS) limit. One family consists {\it on-site symmetric} (vertex-centered) states; these are spatially localized solitary standing waves which are symmetric about any fixed lattice site. The other spatially localized states are {\it off-site symmetric}. Depending on the spatial dimension, these may be bond-centered, cell-centered, or face-centered. Finally, we show that the energy difference among distinct states of the same frequency is exponentially small with respect to a natural parameter. This provides a rigorous bound for the so-called {\it Peierls-Nabarro} energy barrier.
△ Less
Submitted 3 August, 2015; v1 submitted 15 May, 2014;
originally announced May 2014.
