Showing 1–50 of 80 results for author: Petersen, S

Debiasing Cardiac Imaging with Controlled Latent Diffusion Models

Authors: Grzegorz Skorupko, Richard Osuala, Zuzanna Szafranowska, Kaisar Kushibar, Nay Aung, Steffen E Petersen, Karim Lekadir, Polyxeni Gkontra

Abstract: The progress in deep learning solutions for disease diagnosis and prognosis based on cardiac magnetic resonance imaging is hindered by highly imbalanced and biased training data. To address this issue, we propose a method to alleviate imbalances inherent in datasets through the generation of synthetic data based on sensitive attributes such as sex, age, body mass index, and health condition. We ad… ▽ More The progress in deep learning solutions for disease diagnosis and prognosis based on cardiac magnetic resonance imaging is hindered by highly imbalanced and biased training data. To address this issue, we propose a method to alleviate imbalances inherent in datasets through the generation of synthetic data based on sensitive attributes such as sex, age, body mass index, and health condition. We adopt ControlNet based on a denoising diffusion probabilistic model to condition on text assembled from patient metadata and cardiac geometry derived from segmentation masks using a large-cohort study, specifically, the UK Biobank. We assess our method by evaluating the realism of the generated images using established quantitative metrics. Furthermore, we conduct a downstream classification task aimed at debiasing a classifier by rectifying imbalances within underrepresented groups through synthetically generated samples. Our experiments demonstrate the effectiveness of the proposed approach in mitigating dataset imbalances, such as the scarcity of younger patients or individuals with normal BMI level suffering from heart failure. This work represents a major step towards the adoption of synthetic data for the development of fair and generalizable models for medical classification tasks. Notably, we conduct all our experiments using a single, consumer-level GPU to highlight the feasibility of our approach within resource-constrained environments. Our code is available at https://github.com/faildeny/debiasing-cardiac-mri. △ Less

Submitted 28 March, 2024; originally announced March 2024.

The radial variation of the LMC-induced reflex motion of the Milky Way disc observed in the stellar halo

Authors: Rashid Yaaqib, Michael S. Petersen, Jorge Peñarrubia

Abstract: We measure the kinematic signature arising from the Milky Way (MW) disc moving with respect to the outer stellar halo, which is observed as a dipole signal in the kinematics of stellar halo tracers. We quantify how the reflex motion varies as a function of Galactocentric distance, finding that (i) the amplitude of the dipole signal increases as a function of radius, and (ii) the direction moves ac… ▽ More We measure the kinematic signature arising from the Milky Way (MW) disc moving with respect to the outer stellar halo, which is observed as a dipole signal in the kinematics of stellar halo tracers. We quantify how the reflex motion varies as a function of Galactocentric distance, finding that (i) the amplitude of the dipole signal increases as a function of radius, and (ii) the direction moves across the sky. We compare the reflex motion signal against a compilation of published models that follow the MW--LMC interaction. These models show a similar trend of increasing amplitude of the reflex motion as a function of distance, but they do not reproduce the direction of the disc motion with respect to the stellar halo well. We also report mean motions for the stellar halo as a function of distance, finding radial compression in the outer halo and nonzero prograde rotation at all radii. The observed compression signal is also present in MW--LMC models, but the rotation is not, which suggests that the latter is not induced by the LMC. We extensively validate our technique to measure reflex motion against idealised tests. We discuss prospects for directly constraining the mass and orbital history of the LMC through the impact on the motion of the MW stellar disc, and how the modelling of the reflex motion can be improved as more and better data become available. △ Less

Submitted 20 February, 2024; v1 submitted 16 February, 2024; originally announced February 2024.

Comments: 14 pages, 11 figures

Disentangling the Galaxy's Gordian knot: evidence from $APOGEE-Gaia$ for a knotted and slower bar in the Milky Way

Authors: Danny Horta, Michael S. Petersen, Jorge Peñarrubia

Abstract: The inner $\sim5$ kiloparsec (kpc) region of the Milky Way is complex. Unravelling the evolution of the Galaxy requires precise understanding of the formation of this region. We report a study focused on disentangling the inner Galaxy ($r < 5$ kpc) using the measured positions, velocities, and element abundance ratios of red giant stars from the $APOGEE-Gaia$ surveys. After removing the stellar ha… ▽ More The inner $\sim5$ kiloparsec (kpc) region of the Milky Way is complex. Unravelling the evolution of the Galaxy requires precise understanding of the formation of this region. We report a study focused on disentangling the inner Galaxy ($r < 5$ kpc) using the measured positions, velocities, and element abundance ratios of red giant stars from the $APOGEE-Gaia$ surveys. After removing the stellar halo, inner Galaxy populations can be grouped into three main components based on their angular momentum: bar, disc, and a previously unreported ``knot'' component. The knot has a spheroidal shape, is concentrated in the inner $\sim1.5$ kpc, is comprised of stars on nearly-radial orbits, and contains stars with super-solar [Fe/H] element abundances. The chemical compositions of the knot are qualitatively similar to the Galactic bar and inner disc, suggestive that these three populations share a common genesis; the chemical/dynamic properties of the knot suggest it could constitute a classical bulge formed via secular evolution. Moreover, our results show that the bar is more slowly rotating than previously thought, with a pattern speed of $Ω_{\mathrm{bar}}=24\pm3$ km s$^{-1}$ kpc$^{-1}$. This new estimate suggests that the influence of the bar extends beyond the solar radius, with $R_{\mathrm{CR}}\sim9.4-9.8$ kpc, depending on the adopted Milky Way rotation curve; it also suggests a ratio of corotation to bar length of $\mathcal{R}\sim1.8-2$. Our findings help place constraints on the formation and evolution of inner Galaxy populations, and directly constrain dynamical studies of the Milky Way bar and stars in the solar neighbourhood. △ Less

Submitted 12 February, 2024; originally announced February 2024.

Comments: Under review in MNRAS. 15 pages, 10 figures, and 2 tables

Action and energy clustering of stellar streams in deforming Milky Way dark matter haloes

Authors: Richard A. N. Brooks, Jason L. Sanders, Sophia Lilleengen, Michael S. Petersen, Andrew Pontzen

Abstract: We investigate the non-adiabatic effect of time-dependent deformations in the Milky Way (MW) halo potential on stellar streams. Specifically, we consider the MW's response to the infall of the Large Magellanic Cloud (LMC) and how this impacts our ability to recover the spherically averaged MW mass profile from observation using stream actions. Previously, action clustering methods have only been a… ▽ More We investigate the non-adiabatic effect of time-dependent deformations in the Milky Way (MW) halo potential on stellar streams. Specifically, we consider the MW's response to the infall of the Large Magellanic Cloud (LMC) and how this impacts our ability to recover the spherically averaged MW mass profile from observation using stream actions. Previously, action clustering methods have only been applied to static or adiabatic MW systems to constrain the properties of the host system. We use a time-evolving MW--LMC simulation described by basis function expansions. We find that for streams with realistic observational uncertainties on shorter orbital periods and without close encounters with the LMC, e.g. GD-1, the radial action distribution is sufficiently clustered to locally recover the spherical MW mass profile across the stream radial range within a 2 sigma confidence interval determined using a Fisher information approach. For streams with longer orbital periods and close encounters with the LMC, e.g. Orphan-Chenab (OC), the radial action distribution disperses as the MW halo has deformed non-adiabatically. Hence, for OC streams generated in potentials that include a MW halo with any deformations, action clustering methods will fail to recover the spherical mass profile within a 2 sigma uncertainty. Finally, we investigate whether the clustering of stream energies can provide similar constraints. Surprisingly, we find for OC-like streams, the recovered spherically averaged mass profiles demonstrate less sensitivity to the time-dependent deformations in the potential. △ Less

Submitted 21 June, 2024; v1 submitted 22 January, 2024; originally announced January 2024.

Comments: MNRAS accepted. 17 pages, 8 figures

Finiteness properties of torsion fields of abelian varieties

Authors: Wojciech Gajda, Sebastian Petersen

Abstract: Let $A$ be an abelian variety defined over a field $K.$ We study finite generation properties of the profinite group $\mathrm{Gal}(Ω/K)$ and of certain closed normal subgroups thereof, where $Ω$ is the torsion field of $A$ over $K$. In fact, we establish more general finite generation properties for monodromy groups attached to smooth projective varieties via étale cohomology. We apply this in ord… ▽ More Let $A$ be an abelian variety defined over a field $K.$ We study finite generation properties of the profinite group $\mathrm{Gal}(Ω/K)$ and of certain closed normal subgroups thereof, where $Ω$ is the torsion field of $A$ over $K$. In fact, we establish more general finite generation properties for monodromy groups attached to smooth projective varieties via étale cohomology. We apply this in order to give an independent proof and generalizations of a recent result of Checcoli and Dill about small exponent subfields of $Ω/K$ in the number field case. We also give an application of our finite generation results in the realm of permanence principles for varieties with the weak Hilbert property. △ Less

Submitted 1 July, 2024; v1 submitted 11 January, 2024; originally announced January 2024.

Comments: 17 pages. Slightly improved version now treating representations attached to 'etale cohomology as wee

MSC Class: 15K05; 14K15; 11G10

Hilbert properties under base change in small extensions

Authors: Lior Bary-Soroker, Arno Fehm, Sebastian Petersen

Abstract: We study the preservation of the Hilbert property and of the weak Hilbert property under base change in field extensions. In particular we show that these properties are preserved if the extension is finitely generated or Galois with finitely generated Galois group, and we also obtain some negative results. We study the preservation of the Hilbert property and of the weak Hilbert property under base change in field extensions. In particular we show that these properties are preserved if the extension is finitely generated or Galois with finitely generated Galois group, and we also obtain some negative results. △ Less

Submitted 24 December, 2023; originally announced December 2023.

MSC Class: 12E30; 12E25; 14G05; 11G05

Predicting the linear response of self-gravitating stellar spheres and discs with LinearResponse.jl

Authors: Michael S. Petersen, Mathieu Roule, Jean-Baptiste Fouvry, Christophe Pichon, Kerwann Tep

Abstract: We present LinearResponse.jl, an efficient, versatile public library written in julia to compute the linear response of self-gravitating (3D spherically symmetric) stellar spheres and (2D axisymmetric razor-thin) discs. LinearResponse.jl can scan the whole complex frequency plane, probing unstable, neutral and (weakly) damped modes. Given a potential model and a distribution function, this numeric… ▽ More We present LinearResponse.jl, an efficient, versatile public library written in julia to compute the linear response of self-gravitating (3D spherically symmetric) stellar spheres and (2D axisymmetric razor-thin) discs. LinearResponse.jl can scan the whole complex frequency plane, probing unstable, neutral and (weakly) damped modes. Given a potential model and a distribution function, this numerical toolbox estimates the modal frequencies as well as the shapes of individual modes. The libraries are validated against a combination of previous results for the spherical isochrone model and Mestel discs, and new simulations for the spherical Plummer model. Beyond linear response theory, the realm of applications of LinearResponse.jl also extends to the kinetic theory of self-gravitating systems through a modular interface. △ Less

Submitted 17 November, 2023; originally announced November 2023.

Comments: Software available at https://github.com/michael-petersen/LinearResponse.jl

FUTURE-AI: International consensus guideline for trustworthy and deployable artificial intelligence in healthcare

Authors: Karim Lekadir, Aasa Feragen, Abdul Joseph Fofanah, Alejandro F Frangi, Alena Buyx, Anais Emelie, Andrea Lara, Antonio R Porras, An-Wen Chan, Arcadi Navarro, Ben Glocker, Benard O Botwe, Bishesh Khanal, Brigit Beger, Carol C Wu, Celia Cintas, Curtis P Langlotz, Daniel Rueckert, Deogratias Mzurikwao, Dimitrios I Fotiadis, Doszhan Zhussupov, Enzo Ferrante, Erik Meijering, Eva Weicken, Fabio A González , et al. (95 additional authors not shown)

Abstract: Despite major advances in artificial intelligence (AI) for medicine and healthcare, the deployment and adoption of AI technologies remain limited in real-world clinical practice. In recent years, concerns have been raised about the technical, clinical, ethical and legal risks associated with medical AI. To increase real world adoption, it is essential that medical AI tools are trusted and accepted… ▽ More Despite major advances in artificial intelligence (AI) for medicine and healthcare, the deployment and adoption of AI technologies remain limited in real-world clinical practice. In recent years, concerns have been raised about the technical, clinical, ethical and legal risks associated with medical AI. To increase real world adoption, it is essential that medical AI tools are trusted and accepted by patients, clinicians, health organisations and authorities. This work describes the FUTURE-AI guideline as the first international consensus framework for guiding the development and deployment of trustworthy AI tools in healthcare. The FUTURE-AI consortium was founded in 2021 and currently comprises 118 inter-disciplinary experts from 51 countries representing all continents, including AI scientists, clinicians, ethicists, and social scientists. Over a two-year period, the consortium defined guiding principles and best practices for trustworthy AI through an iterative process comprising an in-depth literature review, a modified Delphi survey, and online consensus meetings. The FUTURE-AI framework was established based on 6 guiding principles for trustworthy AI in healthcare, i.e. Fairness, Universality, Traceability, Usability, Robustness and Explainability. Through consensus, a set of 28 best practices were defined, addressing technical, clinical, legal and socio-ethical dimensions. The recommendations cover the entire lifecycle of medical AI, from design, development and validation to regulation, deployment, and monitoring. FUTURE-AI is a risk-informed, assumption-free guideline which provides a structured approach for constructing medical AI tools that will be trusted, deployed and adopted in real-world practice. Researchers are encouraged to take the recommendations into account in proof-of-concept stages to facilitate future translation towards clinical practice of medical AI. △ Less

Submitted 8 July, 2024; v1 submitted 11 August, 2023; originally announced September 2023.

ACM Class: I.2.0; I.4.0; I.5.0

A Thom Spectrum Model for $C_2$-Integral Brown--Gitler Spectra

Authors: Guchuan Li, Sarah Petersen, Elizabeth Ellen Tatum

Abstract: A Thom spectrum model for a $C_2$-equivariant analogue of integral Brown--Gitler spectra is established and shown to have a multiplicative property. The $C_2$-equivariant spectra constructed enjoy properties analogous to classical nonequivariant integral Brown--Gitler spectra and thus may prove useful for producing $C_2$-equivariant analogues of splittings of… ▽ More A Thom spectrum model for a $C_2$-equivariant analogue of integral Brown--Gitler spectra is established and shown to have a multiplicative property. The $C_2$-equivariant spectra constructed enjoy properties analogous to classical nonequivariant integral Brown--Gitler spectra and thus may prove useful for producing $C_2$-equivariant analogues of splittings of $BP \langle 1 \rangle \wedge BP \langle 1 \rangle$ and $bo \wedge bo.$ △ Less

Submitted 5 August, 2024; v1 submitted 24 August, 2023; originally announced August 2023.

Comments: Revision clarifies localizations and completions, 15 pages, comments welcome

Report number: MPIM-Bonn-2022 MSC Class: 55P42; 55P91; 55P92

RomAndromeda: The Roman Survey of the Andromeda Halo

Authors: Arjun Dey, Joan Najita, Carrie Filion, Jiwon Jesse Han, Sarah Pearson, Rosemary Wyse, Adrien C. R. Thob, Borja Anguiano, Miranda Apfel, Magda Arnaboldi, Eric F. Bell, Leandro Beraldo e Silva, Gurtina Besla, Aparajito Bhattacharya, Souradeep Bhattacharya, Vedant Chandra, Yumi Choi, Michelle L. M. Collins, Emily C. Cunningham, Julianne J. Dalcanton, Ivanna Escala, Hayden R. Foote, Annette M. N. Ferguson, Benjamin J. Gibson, Oleg Y. Gnedin , et al. (28 additional authors not shown)

Abstract: As our nearest large neighbor, the Andromeda Galaxy provides a unique laboratory for investigating galaxy formation and the distribution and substructure properties of dark matter in a Milky Way-like galaxy. Here, we propose an initial 2-epoch ($Δt\approx 5$yr), 2-band Roman survey of the entire halo of Andromeda, covering 500 square degrees, which will detect nearly every red giant star in the ha… ▽ More As our nearest large neighbor, the Andromeda Galaxy provides a unique laboratory for investigating galaxy formation and the distribution and substructure properties of dark matter in a Milky Way-like galaxy. Here, we propose an initial 2-epoch ($Δt\approx 5$yr), 2-band Roman survey of the entire halo of Andromeda, covering 500 square degrees, which will detect nearly every red giant star in the halo (10$σ$ detection in F146, F062 of 26.5, 26.1AB mag respectively) and yield proper motions to $\sim$25 microarcsec/year (i.e., $\sim$90 km/s) for all stars brighter than F146 $\approx 23.6$ AB mag (i.e., reaching the red clump stars in the Andromeda halo). This survey will yield (through averaging) high-fidelity proper motions for all satellites and compact substructures in the Andromeda halo and will enable statistical searches for clusters in chemo-dynamical space. Adding a third epoch during the extended mission will improve these proper motions by $\sim t^{-1.5}$, to $\approx 11$ km/s, but this requires obtaining the first epoch in Year 1 of Roman operations. In combination with ongoing and imminent spectroscopic campaigns with ground-based telescopes, this Roman survey has the potential to yield full 3-d space motions of $>$100,000 stars in the Andromeda halo, including (by combining individual measurements) robust space motions of its entire globular cluster and most of its dwarf galaxy satellite populations. It will also identify high-velocity stars in Andromeda, providing unique information on the processes that create this population. These data offer a unique opportunity to study the immigration history, halo formation, and underlying dark matter scaffolding of a galaxy other than our own. △ Less

Submitted 21 June, 2023; originally announced June 2023.

Comments: Submitted in response to the call for Roman Space Telescope Core Community Survey white papers

Measuring the dynamical length of galactic bars

Authors: Michael S. Petersen, Martin D. Weinberg, Neal Katz

Abstract: We define a physically-motivated measure for galactic bar length, called the dynamical length. The dynamical length of the bar corresponds to the radial extent of the orbits that are the backbone supporting the bar feature. We propose a direct observational technique using integral field unit spectroscopy to measure it. Identifying these orbits and using the dynamical length is a more faithful tra… ▽ More We define a physically-motivated measure for galactic bar length, called the dynamical length. The dynamical length of the bar corresponds to the radial extent of the orbits that are the backbone supporting the bar feature. We propose a direct observational technique using integral field unit spectroscopy to measure it. Identifying these orbits and using the dynamical length is a more faithful tracer of the secular evolution and influence of the bar. We demonstrate the success of the metric for recovering the maximal bar-parenting orbit in a range of simulations, and to show its promise we perform its measurement on a real galaxy. We also study the difference between traditionally used ellipse fit approaches to determine bar length and the dynamical length proposed here in a wide range of bar-forming N-body simulations of a stellar disc and dark matter halo. We find that ellipse fitting may severely overestimate measurements of the bar length by a factor of 1.5-2.5 relative to the extent of the orbits that are trapped and actually comprise the bar. This bias leads to overestimates of both bar mass and the ratio of corotation radius to bar length, i.e. the bar speed, affecting inferences about the evolution of bars in the real universe. △ Less

Submitted 22 May, 2023; originally announced May 2023.

Comments: Submitted to MNRAS, comments welcome. arXiv admin note: text overlap with arXiv:1903.08203

A Perspective on Explainable Artificial Intelligence Methods: SHAP and LIME

Authors: Ahmed Salih, Zahra Raisi-Estabragh, Ilaria Boscolo Galazzo, Petia Radeva, Steffen E. Petersen, Gloria Menegaz, Karim Lekadir

Abstract: eXplainable artificial intelligence (XAI) methods have emerged to convert the black box of machine learning (ML) models into a more digestible form. These methods help to communicate how the model works with the aim of making ML models more transparent and increasing the trust of end-users into their output. SHapley Additive exPlanations (SHAP) and Local Interpretable Model Agnostic Explanation (L… ▽ More eXplainable artificial intelligence (XAI) methods have emerged to convert the black box of machine learning (ML) models into a more digestible form. These methods help to communicate how the model works with the aim of making ML models more transparent and increasing the trust of end-users into their output. SHapley Additive exPlanations (SHAP) and Local Interpretable Model Agnostic Explanation (LIME) are two widely used XAI methods, particularly with tabular data. In this perspective piece, we discuss the way the explainability metrics of these two methods are generated and propose a framework for interpretation of their outputs, highlighting their weaknesses and strengths. Specifically, we discuss their outcomes in terms of model-dependency and in the presence of collinearity among the features, relying on a case study from the biomedical domain (classification of individuals with or without myocardial infarction). The results indicate that SHAP and LIME are highly affected by the adopted ML model and feature collinearity, raising a note of caution on their usage and interpretation. △ Less

Submitted 17 June, 2024; v1 submitted 3 May, 2023; originally announced May 2023.

Characterizing the contribution of dependent features in XAI methods

Authors: Ahmed Salih, Ilaria Boscolo Galazzo, Zahra Raisi-Estabragh, Steffen E. Petersen, Gloria Menegaz, Petia Radeva

Abstract: Explainable Artificial Intelligence (XAI) provides tools to help understanding how the machine learning models work and reach a specific outcome. It helps to increase the interpretability of models and makes the models more trustworthy and transparent. In this context, many XAI methods were proposed being SHAP and LIME the most popular. However, the proposed methods assume that used predictors in… ▽ More Explainable Artificial Intelligence (XAI) provides tools to help understanding how the machine learning models work and reach a specific outcome. It helps to increase the interpretability of models and makes the models more trustworthy and transparent. In this context, many XAI methods were proposed being SHAP and LIME the most popular. However, the proposed methods assume that used predictors in the machine learning models are independent which in general is not necessarily true. Such assumption casts shadows on the robustness of the XAI outcomes such as the list of informative predictors. Here, we propose a simple, yet useful proxy that modifies the outcome of any XAI feature ranking method allowing to account for the dependency among the predictors. The proposed approach has the advantage of being model-agnostic as well as simple to calculate the impact of each predictor in the model in presence of collinearity. △ Less

Submitted 4 April, 2023; originally announced April 2023.

Comments: 17 pages, 5 tables

Correlating Changes in Spot Filling Factors with Stellar Rotation: The Case of LkCa 4

Authors: Facundo Pérez Paolino, Jeffrey S. Bary, Michael S. Petersen, Kimberly Ward-Duong, Benjamin M. Tofflemire, Katherine B. Follette, Heidi Mach

Abstract: We present a multi-epoch spectroscopic study of LkCa 4, a heavily spotted non-accreting T Tauri star. Using SpeX at NASA's Infrared Telescope Facility (IRTF), 12 spectra were collected over five consecutive nights, spanning $\approx$ 1.5 stellar rotations. Using the IRTF SpeX Spectral Library, we constructed empirical composite models of spotted stars by combining a warmer (photosphere) standard s… ▽ More We present a multi-epoch spectroscopic study of LkCa 4, a heavily spotted non-accreting T Tauri star. Using SpeX at NASA's Infrared Telescope Facility (IRTF), 12 spectra were collected over five consecutive nights, spanning $\approx$ 1.5 stellar rotations. Using the IRTF SpeX Spectral Library, we constructed empirical composite models of spotted stars by combining a warmer (photosphere) standard star spectrum with a cooler (spot) standard weighted by the spot filling factor, $f_{spot}$. The best-fit models spanned two photospheric component temperatures, $T_{phot}$ = 4100 K (K7V) and 4400 K (K5V), and one spot component temperature, $T_{spot}$ = 3060 K (M5V) with an $A_V$ of 0.3. We find values of $f_{spot}$ to vary between 0.77 and 0.94 with an average uncertainty of $\sim$0.04. The variability of $f_{spot}$ is periodic and correlates with its 3.374 day rotational period. Using a mean value for $f^{mean}_{spot}$ to represent the total spot coverage, we calculated spot corrected values for $T_{eff}$ and $L_\star$. Placing these values alongside evolutionary models developed for heavily spotted young stars, we infer mass and age ranges of 0.45-0.6 $M_\odot$ and 0.50-1.25 Myr, respectively. These inferred values represent a twofold increase in the mass and a twofold decrease in the age as compared to standard evolutionary models. Such a result highlights the need for constraining the contributions of cool and warm regions of young stellar atmospheres when estimating $T_{eff}$ and $L_\star$ to infer masses and ages as well as the necessity for models to account for the effects of these regions on the early evolution of low-mass stars. △ Less

Submitted 2 March, 2023; originally announced March 2023.

Comments: 21 pages, 9 Figures; Accepted for publication in ApJ

Dynamical Data Mining Captures Disc-Halo Couplings that Structure Galaxies

Authors: Alexander Johnson, Michael S. Petersen, Kathryn V. Johnston, Martin D. Weinberg

Abstract: Studying coupling between different galactic components is a challenging problem in galactic dynamics. Using basis function expansions (BFEs) and multichannel singular spectrum analysis (mSSA) as a means of dynamical data mining, we discover evidence for two multi-component disc-halo dipole modes in a Milky-Way-like simulated galaxy. One of the modes grows throughout the simulation, while the othe… ▽ More Studying coupling between different galactic components is a challenging problem in galactic dynamics. Using basis function expansions (BFEs) and multichannel singular spectrum analysis (mSSA) as a means of dynamical data mining, we discover evidence for two multi-component disc-halo dipole modes in a Milky-Way-like simulated galaxy. One of the modes grows throughout the simulation, while the other decays throughout the simulation. The multi-component disc-halo modes are driven primarily by the halo, and have implications for the structural evolution of galaxies, including observations of lopsidedness and other non-axisymmetric structure. In our simulation, the modes create surface density features up to 10 per cent relative to the equilibrium model stellar disc. While the simulated galaxy was constructed to be in equilibrium, BFE+mSSA also uncovered evidence of persistent periodic signals incited by aphysical initial conditions disequilibrium, including rings and weak two-armed spirals, both at the 1 per cent level. The method is sensitive to distinct evolutionary features at and even below the 1 per cent level of surface density variation. The use of mSSA produced clean signals for both modes and disequilibrium, efficiently removing variance owing to estimator noise from the input BFE time series. The discovery of multi-component halo-disc modes is strong motivation for application of BFE+mSSA to the rich zoo of dynamics of multi-component interacting galaxies. △ Less

Submitted 5 January, 2023; originally announced January 2023.

Comments: Submitted to MNRAS, comments welcome

The safe operating space for greenhouse gas emissions

Authors: Steffen Petersen, Morten W. Ryberg, Morten Birkved

Abstract: The Planetary Boundary for Climate Change has been surpassed, and humanity must therefore decide on a pathway back to the safe operating space below the Planetary boundaries to minimise the risk of deleterious or even catastrophic environmental change at continental to global scales. However, the control variables used in the concept do not link to current indicators quantified in life-cycle asses… ▽ More The Planetary Boundary for Climate Change has been surpassed, and humanity must therefore decide on a pathway back to the safe operating space below the Planetary boundaries to minimise the risk of deleterious or even catastrophic environmental change at continental to global scales. However, the control variables used in the concept do not link to current indicators quantified in life-cycle assessments. Translating the Planetary Boundaries into maximum scores for life-cycle assessment indicators is therefore necessary to enable the assessment of whether life-cycle environmental impacts of e.g. products and services respect the Planetary Boundaries. This paper investigates two different approaches for translating the Planetary Boundary for Climate Change into estimations of maximum emission of greenhouse gasses expressed as annual emissions of CO2-equivalents, a commonly used Climate Change indicator in life-cycle assessments. The estimations may serve as an end goal benchmark for the life-cycle environmental impact of products and services in proposed pathways and roadmaps for the regeneration back to the Planetary Boundary for Climate Change. △ Less

Submitted 31 August, 2022; originally announced September 2022.

Comments: 12 pages, 2 figures, 3 tables

How do the dynamics of the Milky Way -- Large Magellanic Cloud system affect gamma-ray constraints on particle dark matter?

Authors: Christopher Eckner, Francesca Calore, Denis Erkal, Sophia Lilleengen, Michael S. Petersen

Abstract: Previous studies on astrophysical dark matter (DM) constraints have all assumed that the Milky Way's (MW) DM halo can be modelled in isolation. However, recent work suggests that the MW's largest dwarf satellite, the Large Magellanic Cloud (LMC), has a mass of 10-20$\%$ that of the MW and is currently merging with our Galaxy. As a result, the DM haloes of the MW and LMC are expected to be strongly… ▽ More Previous studies on astrophysical dark matter (DM) constraints have all assumed that the Milky Way's (MW) DM halo can be modelled in isolation. However, recent work suggests that the MW's largest dwarf satellite, the Large Magellanic Cloud (LMC), has a mass of 10-20$\%$ that of the MW and is currently merging with our Galaxy. As a result, the DM haloes of the MW and LMC are expected to be strongly deformed. We here address and quantify the impact of the dynamical response caused by the passage of the LMC through the MW on the prospects for indirect DM searches. Utilising a set of state-of-the-art numerical simulations of the evolution of the MW-LMC system, we derive the DM distribution in both galaxies at the present time based on the Basis Function Expansion formalism. Consequently, we build $J$-factor all-sky maps of the MW-LMC system in order to study the impact of the LMC passage on gamma-ray indirect searches for thermally produced DM annihilating in the outer MW halo as well as within the LMC halo standalone. We conduct a detailed analysis of 12 years of Fermi-LAT data that incorporates various large-scale gamma-ray emission components and we quantify the systematic uncertainty associated with the imperfect knowledge of the astrophysical gamma-ray sources. We find that the dynamical response caused by the LMC passage can alter the constraints on the velocity-averaged annihilation cross section for weak scale particle DM at a level comparable to the existing observational uncertainty of the MW halo's density profile and total mass. △ Less

Submitted 18 January, 2023; v1 submitted 5 August, 2022; originally announced August 2022.

Comments: 21 pages (15 - main body; 6 - appendices), 10 figures; text corresponds to the version published in MNRAS

Report number: LAPTH-038/22

Local to global principles for homomorphisms of abelian schemes

Authors: Wojciech Gajda, Sebastian Petersen

Abstract: Let $A$ and $B$ be abelian varieties defined over the function field $k(S)$ of a smooth algebraic variety $S/k.$ We establish criteria, in terms of restriction maps to subvarieties of $S,$ for existence of various important classes of $k(S)$-homomorphisms from $A$ to $B,$ e.g., for existence of $k(S)$-isogenies. Our main tools consist of Hilbertianity methods, Tate conjecture as proven by Tate, Za… ▽ More Let $A$ and $B$ be abelian varieties defined over the function field $k(S)$ of a smooth algebraic variety $S/k.$ We establish criteria, in terms of restriction maps to subvarieties of $S,$ for existence of various important classes of $k(S)$-homomorphisms from $A$ to $B,$ e.g., for existence of $k(S)$-isogenies. Our main tools consist of Hilbertianity methods, Tate conjecture as proven by Tate, Zarhin and Faltings, and of the minuscule weights conjecture of Zarhin in the case, when the base field is finite. △ Less

Submitted 11 April, 2023; v1 submitted 19 July, 2022; originally announced July 2022.

Comments: Revised version from April 2023

The $H \underline{\mathbb{F}}_2$-homology of $C_2$-equivariant Eilenberg-MacLane spaces

Authors: Sarah Petersen

Abstract: We extend Ravenel-Wilson Hopf ring techniques to $C_2$-equivariant homotopy theory. Our main application and motivation is a computation of the $RO(C_2)$-graded homology of $C_2$-equivariant Eilenberg-MacLane spaces. The result we obtain for $C_2$-equivariant Eilenberg-MacLane spaces associated to the constant Mackey functor $\underline{\mathbb{F}}_2$ gives a $C_2$-equivariant analogue of the clas… ▽ More We extend Ravenel-Wilson Hopf ring techniques to $C_2$-equivariant homotopy theory. Our main application and motivation is a computation of the $RO(C_2)$-graded homology of $C_2$-equivariant Eilenberg-MacLane spaces. The result we obtain for $C_2$-equivariant Eilenberg-MacLane spaces associated to the constant Mackey functor $\underline{\mathbb{F}}_2$ gives a $C_2$-equivariant analogue of the classical computation due to Serre. We also investigate a twisted bar spectral sequence computing the homology of these equivariant Eilenberg-MacLane spaces and suggest the existence of another twisted bar spectral sequence with $E^2$-page given in terms of a twisted Tor functor. △ Less

Submitted 28 September, 2022; v1 submitted 16 June, 2022; originally announced June 2022.

Comments: Extends prior computational result from some (Theorems 5.2 and 6.4) to all $C_2$-equivariant Eilenberg-MacLane spaces associated to the constant Mackey functor $\underline{\mathbb{F}}_2$ (Theorem 6.3) , 31 pages, 10 figures, Comments welcome

Report number: MPIM-Bonn-2022 MSC Class: 55P91 (Primary) 55P20; 55N91 (Secondary)

An AI tool for automated analysis of large-scale unstructured clinical cine CMR databases

Authors: Jorge Mariscal-Harana, Clint Asher, Vittoria Vergani, Maleeha Rizvi, Louise Keehn, Raymond J. Kim, Robert M. Judd, Steffen E. Petersen, Reza Razavi, Andrew King, Bram Ruijsink, Esther Puyol-Antón

Abstract: Artificial intelligence (AI) techniques have been proposed for automating analysis of short axis (SAX) cine cardiac magnetic resonance (CMR), but no CMR analysis tool exists to automatically analyse large (unstructured) clinical CMR datasets. We develop and validate a robust AI tool for start-to-end automatic quantification of cardiac function from SAX cine CMR in large clinical databases. Our pip… ▽ More Artificial intelligence (AI) techniques have been proposed for automating analysis of short axis (SAX) cine cardiac magnetic resonance (CMR), but no CMR analysis tool exists to automatically analyse large (unstructured) clinical CMR datasets. We develop and validate a robust AI tool for start-to-end automatic quantification of cardiac function from SAX cine CMR in large clinical databases. Our pipeline for processing and analysing CMR databases includes automated steps to identify the correct data, robust image pre-processing, an AI algorithm for biventricular segmentation of SAX CMR and estimation of functional biomarkers, and automated post-analysis quality control to detect and correct errors. The segmentation algorithm was trained on 2793 CMR scans from two NHS hospitals and validated on additional cases from this dataset (n=414) and five external datasets (n=6888), including scans of patients with a range of diseases acquired at 12 different centres using CMR scanners from all major vendors. Median absolute errors in cardiac biomarkers were within the range of inter-observer variability: <8.4mL (left ventricle volume), <9.2mL (right ventricle volume), <13.3g (left ventricular mass), and <5.9% (ejection fraction) across all datasets. Stratification of cases according to phenotypes of cardiac disease and scanner vendors showed good performance across all groups. We show that our proposed tool, which combines image pre-processing steps, a domain-generalisable AI algorithm trained on a large-scale multi-domain CMR dataset and quality control steps, allows robust analysis of (clinical or research) databases from multiple centres, vendors, and cardiac diseases. This enables translation of our tool for use in fully-automated processing of large multi-centre databases. △ Less

Submitted 5 July, 2023; v1 submitted 15 June, 2022; originally announced June 2022.

Comments: Accepted at EHJ Digital Health; Bram Ruijsink and Esther Puyol-Antón are shared last authors

Dynamically constraining the length of the Milky Way bar

Authors: Madeline Lucey, Sarah Pearson, Jason A. S. Hunt, Keith Hawkins, Melissa Ness, Michael S. Petersen, Adrian M. Price-Whelan, Martin D. Weinberg

Abstract: We present a novel method for constraining the length of the Galactic bar using 6D phase space information to directly integrate orbits. We define a pseudo-length for the Galactic bar, named $R_{Freq}$, based on the maximal extent of trapped bar orbits. We find the $R_{Freq}$ measured from orbits is consistent with the $R_{Freq}$ of the assumed potential only when the length of the bar and pattern… ▽ More We present a novel method for constraining the length of the Galactic bar using 6D phase space information to directly integrate orbits. We define a pseudo-length for the Galactic bar, named $R_{Freq}$, based on the maximal extent of trapped bar orbits. We find the $R_{Freq}$ measured from orbits is consistent with the $R_{Freq}$ of the assumed potential only when the length of the bar and pattern speed of said potential is similar to the model from which the initial phase-space coordinates of the orbits are derived. Therefore, one can measure the model's or the Milky Way's bar length from 6D phase-space coordinates by determining which assumed potential leads to a self-consistent measured $R_{Freq}$. When we apply this method to $\approx$210,000 stars in APOGEE DR17 and $Gaia$ eDR3 data, we find a consistent result only for potential models with a dynamical bar length of $\approx$3.5 kpc. We find the Milky Way's trapped bar orbits extend out to only $\approx$3.5 kpc, but there is also an overdensity of stars at the end of the bar out to 4.8 kpc which could be related to an attached spiral arm. We also find that the measured orbital structure of the bar is strongly dependent on the properties of the assumed potential. △ Less

Submitted 2 February, 2023; v1 submitted 3 June, 2022; originally announced June 2022.

Comments: 15 pages, 8 figures, 2 tables, accepted to MNRAS, comments welcome

Ramified covers of abelian varieties over torsion fields

Authors: Lior Bary-Soroker, Arno Fehm, Sebastian Petersen

Abstract: We study rational points on ramified covers of abelian varieties over certain infinite Galois extensions of $\mathbb{Q}$. In particular, we prove that every elliptic curve $E$ over $\mathbb{Q}$ has the weak Hilbert property of Corvaja-Zannier both over the maximal abelian extension $\mathbb{Q}^{\rm ab}$ of $\mathbb{Q}$, and over the field $\mathbb{Q}(A_{\rm tor})$ obtained by adjoining to… ▽ More We study rational points on ramified covers of abelian varieties over certain infinite Galois extensions of $\mathbb{Q}$. In particular, we prove that every elliptic curve $E$ over $\mathbb{Q}$ has the weak Hilbert property of Corvaja-Zannier both over the maximal abelian extension $\mathbb{Q}^{\rm ab}$ of $\mathbb{Q}$, and over the field $\mathbb{Q}(A_{\rm tor})$ obtained by adjoining to $\mathbb{Q}$ all torsion points of some abelian variety $A$ over $\mathbb{Q}$. △ Less

Submitted 17 August, 2023; v1 submitted 3 June, 2022; originally announced June 2022.

Comments: preliminaries section shortened, minor corrections

MSC Class: 12E30; 12E25; 14G05; 11G10; 14K15

The effect of the deforming dark matter haloes of the Milky Way and the Large Magellanic Cloud on the Orphan-Chenab stream

Authors: Sophia Lilleengen, Michael S. Petersen, Denis Erkal, Jorge Peñarrubia, Sergey E. Koposov, Ting S. Li, Lara R. Cullinane, Alexander P. Ji, Kyler Kuehn, Geraint F. Lewis, Dougal Mackey, Andrew B. Pace, Nora Shipp, Daniel B. Zucker, Joss Bland-Hawthorn, Tariq Hilmi

Abstract: It has recently been shown that the Large Magellanic Cloud (LMC) has a substantial effect on the Milky Way's stellar halo and stellar streams. Here, we explore how deformations of the Milky Way and LMC's dark matter haloes affect stellar streams, and whether these effects are observable. In particular, we focus on the Orphan-Chenab (OC) stream which passes particularly close to the LMC, and spans… ▽ More It has recently been shown that the Large Magellanic Cloud (LMC) has a substantial effect on the Milky Way's stellar halo and stellar streams. Here, we explore how deformations of the Milky Way and LMC's dark matter haloes affect stellar streams, and whether these effects are observable. In particular, we focus on the Orphan-Chenab (OC) stream which passes particularly close to the LMC, and spans a large portion of the Milky Way's halo. We represent the Milky Way--LMC system using basis function expansions that capture their evolution in an $N$-body simulation. We present the properties of this system, such as the evolution of the densities and force fields of each galaxy. The OC stream is evolved in this time-dependent, deforming potential, and we investigate the effects of the various moments of the Milky Way and the LMC. We find that the simulated OC stream is strongly influenced by the deformations of both the Milky Way and the LMC, and that this effect is much larger than current observational errors. In particular, the Milky Way dipole has the biggest impact on the stream, followed by the evolution of the LMC's monopole, and the LMC's quadrupole. Detecting these effects would confirm a key prediction of collisionless, cold dark matter, and would be a powerful test of alternative dark matter and alternative gravity models. △ Less

Submitted 10 November, 2022; v1 submitted 3 May, 2022; originally announced May 2022.

Comments: 17 pages, 10 figures. Accepted for publication in MNRAS

Implications of the Milky Way travel velocity for dynamical mass estimates of the Local Group

Authors: Katie Chamberlain, Adrian M. Price-Whelan, Gurtina Besla, Emily C. Cunningham, Nicolás Garavito-Camargo, Jorge Peñarrubia, Michael S. Petersen

Abstract: The total mass of the Local Group (LG) is a fundamental quantity that enables interpreting the orbits of its constituent galaxies and placing the LG in a cosmological context. One of the few methods that allows inferring the total mass directly is the "Timing Argument," which models the relative orbit of the Milky Way (MW) and M31 in equilibrium. The MW itself is not in equilibrium, a byproduct of… ▽ More The total mass of the Local Group (LG) is a fundamental quantity that enables interpreting the orbits of its constituent galaxies and placing the LG in a cosmological context. One of the few methods that allows inferring the total mass directly is the "Timing Argument," which models the relative orbit of the Milky Way (MW) and M31 in equilibrium. The MW itself is not in equilibrium, a byproduct of its merger history including the recent pericentric passage of the LMC, and recent work has found that the MW disk is moving with a lower bound "travel velocity" of $\sim 32~{\rm km}~{\rm s}^{-1}$ with respect to the outer stellar halo. Previous Timing Argument measurements attempt to account for this non-equilibrium state, but have been restricted to theoretical predictions for the impact of the LMC specifically. In this paper, we quantify the impact of a travel velocity on recovered LG mass estimates using several different compilations of recent kinematic measurements of M31. We find that incorporating the measured value of the travel velocity lowers the inferred LG mass by 10--12\% compared to a static MW halo. Measurements of the travel velocity with more distant tracers could yield even larger values, which would further decrease the inferred LG mass. Therefore, the newly measured travel velocity directly implies a lower LG mass than from a model with a static MW halo and must be considered in future dynamical studies of the Local Volume. △ Less

Submitted 30 December, 2022; v1 submitted 14 April, 2022; originally announced April 2022.

Comments: 14 pages, 3 figures; Published in ApJ; updated to match published version

Journal ref: ApJ 942 18 (2023)

Effects on the local dark matter distribution due to the Large Magellanic Cloud

Authors: Katelin Donaldson, Michael S. Petersen, Jorge Peñarrubia

Abstract: We study the local dark matter distribution in two models for the Milky Way (MW)-Large Magellanic Cloud (LMC) interaction. The effect of the LMC on the local dark matter distribution is dependent on the evolution of the MW-LMC system, such that a static model is insufficient to accurately model the dark matter velocity distribution in the solar neighbourhood. An evolved model boosts local LMC dark… ▽ More We study the local dark matter distribution in two models for the Milky Way (MW)-Large Magellanic Cloud (LMC) interaction. The effect of the LMC on the local dark matter distribution is dependent on the evolution of the MW-LMC system, such that a static model is insufficient to accurately model the dark matter velocity distribution in the solar neighbourhood. An evolved model boosts local LMC dark matter particle velocities by nearly 50%, to a median value of $\approx750$km/s. MW dark matter particles also experience a velocity boost, which we identify as being caused by reflex motion owing to the infall of the LMC. We study the implications of LMC particles in the solar neighbourhood for dark matter detection experiments. Specifically, the directionality of LMC particles is distinguishable from the MW particles, with a difference in the apparent origin centroid location between the MW and LMC particles of $26\pm6 ^\circ$. This unique identifier, along with their high velocities, can be utilised by directional detectors to search for dark matter particles originating in the LMC. △ Less

Submitted 25 March, 2022; v1 submitted 30 November, 2021; originally announced November 2021.

Comments: 6 pages, 3 figures

Tidally stripped halo stars from the Large Magellanic Cloud in the Galactic North

Authors: Michael S. Petersen, Jorge Peñarrubia, Ella Jones

Abstract: We examine whether the Large Magellanic Cloud (LMC) is currently losing its stellar halo to Milky Way (MW) tides. We present a live $N$-body model for the ongoing MW-LMC interaction that predicts a prominent stream of stars tidally stripped from the progenitor LMC. We use this model to define a strategy to search for stripped material in kinematic space. Of the available stellar tracers, we conclu… ▽ More We examine whether the Large Magellanic Cloud (LMC) is currently losing its stellar halo to Milky Way (MW) tides. We present a live $N$-body model for the ongoing MW-LMC interaction that predicts a prominent stream of stars tidally stripped from the progenitor LMC. We use this model to define a strategy to search for stripped material in kinematic space. Of the available stellar tracers, we conclude that samples of RR Lyrae stars provide the highest density of kinematic tracers at present. Using a sample of RR Lyrae stars with Gaia EDR3 astrometry we show that the LMC stellar halo in the Southern Galactic hemisphere extends at least out to $\sim 30^\circ$ from the galaxy centre. In addition, several leading arm candidates are found in the Northern hemisphere as far above the disc plane as $b=+34^\circ$ (at 68$^\circ$ from the LMC). △ Less

Submitted 25 June, 2021; originally announced June 2021.

Comments: Submitted to MNRAS; this version features minor updates. Comments welcome

Fairness in Cardiac MR Image Analysis: An Investigation of Bias Due to Data Imbalance in Deep Learning Based Segmentation

Authors: Esther Puyol-Anton, Bram Ruijsink, Stefan K. Piechnik, Stefan Neubauer, Steffen E. Petersen, Reza Razavi, Andrew P. King

Abstract: The subject of "fairness" in artificial intelligence (AI) refers to assessing AI algorithms for potential bias based on demographic characteristics such as race and gender, and the development of algorithms to address this bias. Most applications to date have been in computer vision, although some work in healthcare has started to emerge. The use of deep learning (DL) in cardiac MR segmentation ha… ▽ More The subject of "fairness" in artificial intelligence (AI) refers to assessing AI algorithms for potential bias based on demographic characteristics such as race and gender, and the development of algorithms to address this bias. Most applications to date have been in computer vision, although some work in healthcare has started to emerge. The use of deep learning (DL) in cardiac MR segmentation has led to impressive results in recent years, and such techniques are starting to be translated into clinical practice. However, no work has yet investigated the fairness of such models. In this work, we perform such an analysis for racial/gender groups, focusing on the problem of training data imbalance, using a nnU-Net model trained and evaluated on cine short axis cardiac MR data from the UK Biobank dataset, consisting of 5,903 subjects from 6 different racial groups. We find statistically significant differences in Dice performance between different racial groups. To reduce the racial bias, we investigated three strategies: (1) stratified batch sampling, in which batch sampling is stratified to ensure balance between racial groups; (2) fair meta-learning for segmentation, in which a DL classifier is trained to classify race and jointly optimized with the segmentation model; and (3) protected group models, in which a different segmentation model is trained for each racial group. We also compared the results to the scenario where we have a perfectly balanced database. To assess fairness we used the standard deviation (SD) and skewed error ratio (SER) of the average Dice values. Our results demonstrate that the racial bias results from the use of imbalanced training data, and that all proposed bias mitigation strategies improved fairness, with the best SD and SER resulting from the use of protected group models. △ Less

Submitted 1 July, 2021; v1 submitted 23 June, 2021; originally announced June 2021.

Comments: MICCAI 2021 conference

Identification of Sagittarius stream members in Angular Momentum space with Gaussian mixture techniques

Authors: Jorge Peñarrubia, Michael S. Petersen

Abstract: This paper uses Gaussian mixture techniques to dissect the Milky Way (MW) stellar halo in angular momentum space. Application to a catalogue of 5389 stars near the plane of the Sagittarius (Sgr) stream with full 6D phase-space coordinates supplied by Gaia EDR3 and SEGUE returns four independent dynamical components. The broadest and most populated corresponds to the `smooth' MW halo. The narrowest… ▽ More This paper uses Gaussian mixture techniques to dissect the Milky Way (MW) stellar halo in angular momentum space. Application to a catalogue of 5389 stars near the plane of the Sagittarius (Sgr) stream with full 6D phase-space coordinates supplied by Gaia EDR3 and SEGUE returns four independent dynamical components. The broadest and most populated corresponds to the `smooth' MW halo. The narrowest and faintest contains 40 stars of the Orphan stream. We find a component with little or no angular momentum likely associated with the GSE substructure. We also identify 925 stars and 7 Globular Clusters with probabilities $>90\%$ to be members of the Sgr stream. Comparison against $N$-body models shows that some of these members trace the continuation of the leading/trailing tails in the Southern/Northern hemispheres. The new detections span $\sim 800^\circ$ on the sky, thus wrapping the Galaxy {\it twice}. △ Less

Submitted 22 June, 2021; originally announced June 2021.

Comments: Submitted to MNRAS Letters. Comments welcome

The NewHorizon Simulation -- To Bar Or Not To Bar

Authors: J. Reddish, K. Kraljic, M. S. Petersen, K. Tep, Y. Dubois, C. Pichon, S. Peirani, F. Bournaud, H. Choi, J. Devriendt, R. Jackson, G. Martin, M. J. Park, M. Volonteri, S. K. Yi

Abstract: We use the NewHorizon simulation to study the redshift evolution of bar properties and fractions within galaxies in the stellar masses range $M_{\star} = 10^{7.25} - 10^{11.4} \ \rm{M}_{\odot}$ over the redshift range $z = 0.25 - 1.3$. We select disc galaxies using stellar kinematics as a proxy for galaxy morphology. We employ two different automated bar detection methods, coupled with visual insp… ▽ More We use the NewHorizon simulation to study the redshift evolution of bar properties and fractions within galaxies in the stellar masses range $M_{\star} = 10^{7.25} - 10^{11.4} \ \rm{M}_{\odot}$ over the redshift range $z = 0.25 - 1.3$. We select disc galaxies using stellar kinematics as a proxy for galaxy morphology. We employ two different automated bar detection methods, coupled with visual inspection, resulting in observable bar fractions of $f_{\rm bar} = 0.070_{-0.012}^{+0.018}$ at $z\sim$ 1.3, decreasing to $f_{\rm bar} = 0.011_{-0.003}^{+0.014}$ at $z\sim$ 0.25. Only one galaxy is visually confirmed as strongly barred in our sample. This bar is hosted by the most massive disk and only survives from $z=1.3$ down to $z=0.7$. Such a low bar fraction, in particular amongst Milky Way-like progenitors, highlights a missing bars problem, shared by literally all cosmological simulations with spatial resolution $<$100 pc to date. The analysis of linear growth rates, rotation curves and derived summary statistics of the stellar, gas and dark matter components suggest that galaxies with stellar masses below $10^{9.5}-10^{10} \ \rm{M}_{\odot}$ in NewHorizon appear to be too dominated by dark matter to form a bar, while more massive galaxies typically have formed large bulges that prevent bar persistence at low redshift. This investigation confirms that the evolution of the bar fraction puts stringent constraints on the assembly history of baryons and dark matter onto galaxies. △ Less

Submitted 24 February, 2022; v1 submitted 4 June, 2021; originally announced June 2021.

Comments: 25 pages, 13 figures, 5 tables, accepted by Monthly Notices of Royal Astronomical Society (MNRAS) on 16/02/22

EXP: N-body integration using basis function expansions

Authors: Michael S. Petersen, Martin D. Weinberg, Neal Katz

Abstract: We present the N-body simulation techniques in EXP. EXP uses empirically-chosen basis functions to expand the potential field of an ensemble of particles. Unlike other basis function expansions, the derived basis functions are adapted to an input mass distribution, enabling accurate expansion of highly non-spherical objects, such as galactic discs. We measure the force accuracy in three models, on… ▽ More We present the N-body simulation techniques in EXP. EXP uses empirically-chosen basis functions to expand the potential field of an ensemble of particles. Unlike other basis function expansions, the derived basis functions are adapted to an input mass distribution, enabling accurate expansion of highly non-spherical objects, such as galactic discs. We measure the force accuracy in three models, one based on a spherical or aspherical halo, one based on an exponential disc, and one based on a bar-based disc model. We find that EXP is as accurate as a direct-summation or tree-based calculation, and in some ways is better, while being considerably less computationally intensive. We discuss optimising the computation of the basis function representation. We also detail numerical improvements for performing orbit integrations, including timesteps. △ Less

Submitted 29 April, 2021; originally announced April 2021.

Comments: Submitted to MNRAS, comments welcome

Estimating Uncertainty in Neural Networks for Cardiac MRI Segmentation: A Benchmark Study

Authors: Matthew Ng, Fumin Guo, Labonny Biswas, Steffen E. Petersen, Stefan K. Piechnik, Stefan Neubauer, Graham Wright

Abstract: Objective: Convolutional neural networks (CNNs) have demonstrated promise in automated cardiac magnetic resonance image segmentation. However, when using CNNs in a large real-world dataset, it is important to quantify segmentation uncertainty and identify segmentations which could be problematic. In this work, we performed a systematic study of Bayesian and non-Bayesian methods for estimating unce… ▽ More Objective: Convolutional neural networks (CNNs) have demonstrated promise in automated cardiac magnetic resonance image segmentation. However, when using CNNs in a large real-world dataset, it is important to quantify segmentation uncertainty and identify segmentations which could be problematic. In this work, we performed a systematic study of Bayesian and non-Bayesian methods for estimating uncertainty in segmentation neural networks. Methods: We evaluated Bayes by Backprop, Monte Carlo Dropout, Deep Ensembles, and Stochastic Segmentation Networks in terms of segmentation accuracy, probability calibration, uncertainty on out-of-distribution images, and segmentation quality control. Results: We observed that Deep Ensembles outperformed the other methods except for images with heavy noise and blurring distortions. We showed that Bayes by Backprop is more robust to noise distortions while Stochastic Segmentation Networks are more resistant to blurring distortions. For segmentation quality control, we showed that segmentation uncertainty is correlated with segmentation accuracy for all the methods. With the incorporation of uncertainty estimates, we were able to reduce the percentage of poor segmentation to 5% by flagging 31--48% of the most uncertain segmentations for manual review, substantially lower than random review without using neural network uncertainty (reviewing 75--78% of all images). Conclusion: This work provides a comprehensive evaluation of uncertainty estimation methods and showed that Deep Ensembles outperformed other methods in most cases. Significance: Neural network uncertainty measures can help identify potentially inaccurate segmentations and alert users for manual review. △ Less

Submitted 30 December, 2022; v1 submitted 31 December, 2020; originally announced December 2020.

Comments: Accepted to IEEE Transactions on Biomedical Engineering. Copyright (c) 2022 IEEE. Personal use of this material is permitted. However, permission to use this material for any other purposes must be obtained from the IEEE by sending an email to [email protected]

Detection of the Milky Way reflex motion due to the Large Magellanic Cloud infall

Authors: Michael S. Petersen, Jorge Peñarrubia

Abstract: The Large Magellanic Cloud is the most massive satellite galaxy of the Milky Way, with an estimated mass exceeding a tenth of the mass of the Milky Way. Just past its closest approach of about 50 kpc, and flying by the Milky Way at an astonishing speed of 327 km/s, the Large Magellanic Cloud can affect our Galaxy in a number of ways, including dislodging the Milky Way disc from the Galactic centre… ▽ More The Large Magellanic Cloud is the most massive satellite galaxy of the Milky Way, with an estimated mass exceeding a tenth of the mass of the Milky Way. Just past its closest approach of about 50 kpc, and flying by the Milky Way at an astonishing speed of 327 km/s, the Large Magellanic Cloud can affect our Galaxy in a number of ways, including dislodging the Milky Way disc from the Galactic centre-of-mass. Here, we report evidence that the Milky Way disc is moving with respect to stellar tracers in the outer halo ($40<r<120$ kpc) at $v_{\rm travel}=32^{+4}_{-4}$ km/s, in the direction $(\ell,b)_{\rm apex}=(56^{+9}_{-9},-34^{+10}_{-9})$ degrees, which points at an earlier location on the LMC trajectory. The resulting reflex motion is detected in the kinematics of outer halo stars and Milky Way satellite galaxies with accurate distances, proper motions and line-of-sight velocities. Our results indicate that dynamical models of our Galaxy cannot neglect gravitational perturbations induced by the Large Magellanic Cloud infall, nor can observations of the stellar halo be treated in a reference frame that does not correct for disc reflex motion. Future spectroscopic surveys of the stellar halo combined with Gaia astrometry will allow for sophisticated modelling of the Large Magellanic Cloud trajectory across the Milky Way, constraining the dark matter distribution in both galaxies with unprecedented detail. △ Less

Submitted 20 November, 2020; originally announced November 2020.

Comments: Author's version of the paper published 23 November 2020 in Nature Astronomy

Using Multichannel Singular Spectrum Analysis to Study Galaxy Dynamics

Authors: Martin D. Weinberg, Michael S. Petersen

Abstract: N-body simulations provide most of our insight into the structure and evolution of galaxies, but our analyses of these are often heuristic and from simple statistics. We propose a method that discovers the dynamics in space and time together by finding the most correlated temporal signals in multiple time series of basis function expansion coefficients and any other data fields of interest. The me… ▽ More N-body simulations provide most of our insight into the structure and evolution of galaxies, but our analyses of these are often heuristic and from simple statistics. We propose a method that discovers the dynamics in space and time together by finding the most correlated temporal signals in multiple time series of basis function expansion coefficients and any other data fields of interest. The method extracts the dominant trends in the spatial variation of the gravitational field along with any additional data fields through time. The mathematics of this method is known as multichannel singular spectrum analysis (M-SSA). In essence, M-SSA is a principal component analysis of the covariance of time series replicates, each lagged successively by some interval. The dominant principal component represents the trend that contains the largest fraction of the correlated signal. The next principal component is orthogonal to the first and contains the next largest fraction, and so on. Using a suite of previously analysed simulations, we find that M-SSA describes bar formation and evolution, including mode coupling and pattern-speed decay. We also analyse a new simulation tailored to study vertical oscillations of the bar using kinematic data. Additionally, and to our surprise, M-SSA uncovered some new dynamics in previously analysed simulations, underscoring the power of this new approach. △ Less

Submitted 16 September, 2020; originally announced September 2020.

Comments: 18 pages, 14 figures, uses pdflatex, submitted to MNRAS

A radiomics approach to analyze cardiac alterations in hypertension

Authors: Irem Cetin, Steffen E. Petersen, Sandy Napel, Oscar Camara, Miguel Angel Gonzalez Ballester, Karim Lekadir

Abstract: Hypertension is a medical condition that is well-established as a risk factor for many major diseases. For example, it can cause alterations in the cardiac structure and function over time that can lead to heart related morbidity and mortality. However, at the subclinical stage, these changes are subtle and cannot be easily captured using conventional cardiovascular indices calculated from clinica… ▽ More Hypertension is a medical condition that is well-established as a risk factor for many major diseases. For example, it can cause alterations in the cardiac structure and function over time that can lead to heart related morbidity and mortality. However, at the subclinical stage, these changes are subtle and cannot be easily captured using conventional cardiovascular indices calculated from clinical cardiac imaging. In this paper, we describe a radiomics approach for identifying intermediate imaging phenotypes associated with hypertension. The method combines feature selection and machine learning techniques to identify the most subtle as well as complex structural and tissue changes in hypertensive subgroups as compared to healthy individuals. Validation based on a sample of asymptomatic hearts that include both hypertensive and non-hypertensive cases demonstrate that the proposed radiomics model is capable of detecting intensity and textural changes well beyond the capabilities of conventional imaging phenotypes, indicating its potential for improved understanding of the longitudinal effects of hypertension on cardiovascular health and disease. △ Less

Submitted 21 July, 2020; originally announced July 2020.

Fully Automated Myocardial Strain Estimation from CMR Tagged Images using a Deep Learning Framework in the UK Biobank

Authors: Edward Ferdian, Avan Suinesiaputra, Kenneth Fung, Nay Aung, Elena Lukaschuk, Ahmet Barutcu, Edd Maclean, Jose Paiva, Stefan K. Piechnik, Stefan Neubauer, Steffen E Petersen, Alistair A. Young

Abstract: Purpose: To demonstrate the feasibility and performance of a fully automated deep learning framework to estimate myocardial strain from short-axis cardiac magnetic resonance tagged images. Methods and Materials: In this retrospective cross-sectional study, 4508 cases from the UK Biobank were split randomly into 3244 training and 812 validation cases, and 452 test cases. Ground truth myocardial lan… ▽ More Purpose: To demonstrate the feasibility and performance of a fully automated deep learning framework to estimate myocardial strain from short-axis cardiac magnetic resonance tagged images. Methods and Materials: In this retrospective cross-sectional study, 4508 cases from the UK Biobank were split randomly into 3244 training and 812 validation cases, and 452 test cases. Ground truth myocardial landmarks were defined and tracked by manual initialization and correction of deformable image registration using previously validated software with five readers. The fully automatic framework consisted of 1) a convolutional neural network (CNN) for localization, and 2) a combination of a recurrent neural network (RNN) and a CNN to detect and track the myocardial landmarks through the image sequence for each slice. Radial and circumferential strain were then calculated from the motion of the landmarks and averaged on a slice basis. Results: Within the test set, myocardial end-systolic circumferential Green strain errors were -0.001 +/- 0.025, -0.001 +/- 0.021, and 0.004 +/- 0.035 in basal, mid, and apical slices respectively (mean +/- std. dev. of differences between predicted and manual strain). The framework reproduced significant reductions in circumferential strain in diabetics, hypertensives, and participants with previous heart attack. Typical processing time was ~260 frames (~13 slices) per second on an NVIDIA Tesla K40 with 12GB RAM, compared with 6-8 minutes per slice for the manual analysis. Conclusions: The fully automated RNNCNN framework for analysis of myocardial strain enabled unbiased strain evaluation in a high-throughput workflow, with similar ability to distinguish impairment due to diabetes, hypertension, and previous heart attack. △ Less

Submitted 15 April, 2020; originally announced April 2020.

Comments: accepted in Radiology Cardiothoracic Imaging

Journal ref: Radiology: Cardiothoracic Imaging 2020; 2(1):e190032

Demonstrating Immersive Media Delivery on 5G Broadcast and Multicast Testing Networks

Authors: De Mi, Joe Eyles, Tero Jokela, Swen Petersen, Roman Odarchenko, Ece Ozturk, Duy-Kha Chau, Tuan Tran, Rory Turnbull, Heikki Kokkinen, Baruch Altman, Menno Bot, Darko Ratkaj, Olaf Renner, David Gomez-Barquero, Jordi Joan Gimenez

Abstract: This work presents eight demonstrators and one showcase developed within the 5G-Xcast project. They experimentally demonstrate and validate key technical enablers for the future of media delivery, associated with multicast and broadcast communication capabilities in 5th Generation (5G). In 5G-Xcast, three existing testbeds: IRT in Munich (Germany), 5GIC in Surrey (UK), and TUAS in Turku (Finland),… ▽ More This work presents eight demonstrators and one showcase developed within the 5G-Xcast project. They experimentally demonstrate and validate key technical enablers for the future of media delivery, associated with multicast and broadcast communication capabilities in 5th Generation (5G). In 5G-Xcast, three existing testbeds: IRT in Munich (Germany), 5GIC in Surrey (UK), and TUAS in Turku (Finland), have been developed into 5G broadcast and multicast testing networks, which enables us to demonstrate our vision of a converged 5G infrastructure with fixed and mobile accesses and terrestrial broadcast, delivering immersive audio-visual media content. Built upon the improved testing networks, the demonstrators and showcase developed in 5G-Xcast show the impact of the technology developed in the project. Our demonstrations predominantly cover use cases belonging to two verticals: Media & Entertainment and Public Warning, which are future 5G scenarios relevant to multicast and broadcast delivery. In this paper, we present the development of these demonstrators, the showcase, and the testbeds. We also provide key findings from the experiments and demonstrations, which not only validate the technical solutions developed in the project, but also illustrate the potential technical impact of these solutions for broadcasters, content providers, operators, and other industries interested in the future immersive media delivery. △ Less

Submitted 1 March, 2020; originally announced April 2020.

Comments: 16 pages, 22 figures, IEEE Trans. Broadcasting

Reflex motion in the Milky Way stellar halo resulting from the Large Magellanic Cloud infall

Authors: Michael S. Petersen, Jorge Peñarrubia

Abstract: We present the results of N-body models of the Milky Way and Large Magellanic Cloud system and study the kinematic reflex motion in the stellar halo owing to the barycentre displacement of the disc. In agreement with previous studies, we find that the Milky Way disc may be moving at 40 km/s relative to the barycentre prior to the Large Magellanic Cloud infall. The resulting reflex motion is visibl… ▽ More We present the results of N-body models of the Milky Way and Large Magellanic Cloud system and study the kinematic reflex motion in the stellar halo owing to the barycentre displacement of the disc. In agreement with previous studies, we find that the Milky Way disc may be moving at 40 km/s relative to the barycentre prior to the Large Magellanic Cloud infall. The resulting reflex motion is visible in tangential velocities of the stellar halo as a simple dipole. The signal is strongest for stars with long dynamical times, identifiable in position-velocity data as stars with large apocentres, whose dynamical memory is still well-represented by the unperturbed Milky Way potential. The signal varies across the sky depending on the stellar tracer and may be up to the same magnitude as the velocity of the disc centre-of-mass, making reflex motion a source of bias for Milky Way potential determinations based on the modeling of stellar streams and/or smooth halo tracers such as blue horizontal branch or RR Lyrae stars. △ Less

Submitted 27 February, 2020; v1 submitted 24 January, 2020; originally announced January 2020.

Ranks of abelian varieties and the full Mordell-Lang conjecture in dimension one

Authors: Arno Fehm, Sebastian Petersen

Abstract: Let $A$ be a non-zero abelian variety over a field $F$ that is not algebraic over a finite field. We prove that the rational rank of the abelian group $A(F)$ is infinite when $F$ is large in the sense of Pop (also called ample). The main ingredient is a deduction of the 1-dimensional case of the relative Mordell-Lang conjecture from a result of Rössler. Let $A$ be a non-zero abelian variety over a field $F$ that is not algebraic over a finite field. We prove that the rational rank of the abelian group $A(F)$ is infinite when $F$ is large in the sense of Pop (also called ample). The main ingredient is a deduction of the 1-dimensional case of the relative Mordell-Lang conjecture from a result of Rössler. △ Less

Submitted 23 December, 2019; originally announced December 2019.

MSC Class: 14K05; 12E30; 14G05

A Radiomics Approach to Computer-Aided Diagnosis with Cardiac Cine-MRI

Authors: Irem Cetin, Gerard Sanroma, Steffen E. Petersen, Sandy Napel, Oscar Camara, Miguel-Angel Gonzalez Ballester, Karim Lekadir

Abstract: Use expert visualization or conventional clinical indices can lack accuracy for borderline classications. Advanced statistical approaches based on eigen-decomposition have been mostly concerned with shape and motion indices. In this paper, we present a new approach to identify CVDs from cine-MRI by estimating large pools of radiomic features (statistical, shape and textural features) encoding rele… ▽ More Use expert visualization or conventional clinical indices can lack accuracy for borderline classications. Advanced statistical approaches based on eigen-decomposition have been mostly concerned with shape and motion indices. In this paper, we present a new approach to identify CVDs from cine-MRI by estimating large pools of radiomic features (statistical, shape and textural features) encoding relevant changes in anatomical and image characteristics due to CVDs. The calculated cine-MRI radiomic features are assessed using sequential forward feature selection to identify the most relevant ones for given CVD classes (e.g. myocardial infarction, cardiomyopathy, abnormal right ventricle). Finally, advanced machine learning is applied to suitably integrate the selected radiomics for final multi-feature classification based on Support Vector Machines (SVMs). The proposed technique was trained and cross-validated using 100 cine-MRI cases corresponding to five different cardiac classes from the ACDC MICCAI 2017 challenge \footnote{https://www.creatis.insa-lyon.fr/Challenge/acdc/index.html}. All cases were correctly classified in this preliminary study, indicating potential of using large-scale radiomics for MRI-based diagnosis of CVDs. △ Less

Submitted 25 September, 2019; originally announced September 2019.

Combining Multi-Sequence and Synthetic Images for Improved Segmentation of Late Gadolinium Enhancement Cardiac MRI

Authors: Víctor M. Campello, Carlos Martín-Isla, Cristian Izquierdo, Steffen E. Petersen, Miguel A. González Ballester, Karim Lekadir

Abstract: Accurate segmentation of the cardiac boundaries in late gadolinium enhancement magnetic resonance images (LGE-MRI) is a fundamental step for accurate quantification of scar tissue. However, while there are many solutions for automatic cardiac segmentation of cine images, the presence of scar tissue can make the correct delineation of the myocardium in LGE-MRI challenging even for human experts. As… ▽ More Accurate segmentation of the cardiac boundaries in late gadolinium enhancement magnetic resonance images (LGE-MRI) is a fundamental step for accurate quantification of scar tissue. However, while there are many solutions for automatic cardiac segmentation of cine images, the presence of scar tissue can make the correct delineation of the myocardium in LGE-MRI challenging even for human experts. As part of the Multi-Sequence Cardiac MR Segmentation Challenge, we propose a solution for LGE-MRI segmentation based on two components. First, a generative adversarial network is trained for the task of modality-to-modality translation between cine and LGE-MRI sequences to obtain extra synthetic images for both modalities. Second, a deep learning model is trained for segmentation with different combinations of original, augmented and synthetic sequences. Our results based on three magnetic resonance sequences (LGE, bSSFP and T2) from 45 different patients show that the multi-sequence model training integrating synthetic images and data augmentation improves in the segmentation over conventional training with real datasets. In conclusion, the accuracy of the segmentation of LGE-MRI images can be improved by using complementary information provided by non-contrast MRI sequences. △ Less

Submitted 13 January, 2020; v1 submitted 3 September, 2019; originally announced September 2019.

Comments: 10 pages, Accepted to MS-CMRSeg Challenge (STACOM 2019), reference added and affiliations updated

Joint Motion Estimation and Segmentation from Undersampled Cardiac MR Image

Authors: Chen Qin, Wenjia Bai, Jo Schlemper, Steffen E. Petersen, Stefan K. Piechnik, Stefan Neubauer, Daniel Rueckert

Abstract: Accelerating the acquisition of magnetic resonance imaging (MRI) is a challenging problem, and many works have been proposed to reconstruct images from undersampled k-space data. However, if the main purpose is to extract certain quantitative measures from the images, perfect reconstructions may not always be necessary as long as the images enable the means of extracting the clinically relevant me… ▽ More Accelerating the acquisition of magnetic resonance imaging (MRI) is a challenging problem, and many works have been proposed to reconstruct images from undersampled k-space data. However, if the main purpose is to extract certain quantitative measures from the images, perfect reconstructions may not always be necessary as long as the images enable the means of extracting the clinically relevant measures. In this paper, we work on jointly predicting cardiac motion estimation and segmentation directly from undersampled data, which are two important steps in quantitatively assessing cardiac function and diagnosing cardiovascular diseases. In particular, a unified model consisting of both motion estimation branch and segmentation branch is learned by optimising the two tasks simultaneously. Additional corresponding fully-sampled images are incorporated into the network as a parallel sub-network to enhance and guide the learning during the training process. Experimental results using cardiac MR images from 220 subjects show that the proposed model is robust to undersampled data and is capable of predicting results that are close to that from fully-sampled ones, while bypassing the usual image reconstruction stage. △ Less

Submitted 20 August, 2019; originally announced August 2019.

Comments: This work is published at MLMIR 2018: Machine Learning for Medical Image Reconstruction

Learning Shape Priors for Robust Cardiac MR Segmentation from Multi-view Images

Authors: Chen Chen, Carlo Biffi, Giacomo Tarroni, Steffen Petersen, Wenjia Bai, Daniel Rueckert

Abstract: Cardiac MR image segmentation is essential for the morphological and functional analysis of the heart. Inspired by how experienced clinicians assess the cardiac morphology and function across multiple standard views (i.e. long- and short-axis views), we propose a novel approach which learns anatomical shape priors across different 2D standard views and leverages these priors to segment the left ve… ▽ More Cardiac MR image segmentation is essential for the morphological and functional analysis of the heart. Inspired by how experienced clinicians assess the cardiac morphology and function across multiple standard views (i.e. long- and short-axis views), we propose a novel approach which learns anatomical shape priors across different 2D standard views and leverages these priors to segment the left ventricular (LV) myocardium from short-axis MR image stacks. The proposed segmentation method has the advantage of being a 2D network but at the same time incorporates spatial context from multiple, complementary views that span a 3D space. Our method achieves accurate and robust segmentation of the myocardium across different short-axis slices (from apex to base), outperforming baseline models (e.g. 2D U-Net, 3D U-Net) while achieving higher data efficiency. Compared to the 2D U-Net, the proposed method reduces the mean Hausdorff distance (mm) from 3.24 to 2.49 on the apical slices, from 2.34 to 2.09 on the middle slices and from 3.62 to 2.76 on the basal slices on the test set, when only 10% of the training data was used. △ Less

Submitted 17 December, 2019; v1 submitted 23 July, 2019; originally announced July 2019.

Comments: 11 pages, 5 figures, accepted at MICCAI 2019, Camera-ready version

Self-Supervised Learning for Cardiac MR Image Segmentation by Anatomical Position Prediction

Authors: Wenjia Bai, Chen Chen, Giacomo Tarroni, Jinming Duan, Florian Guitton, Steffen E. Petersen, Yike Guo, Paul M. Matthews, Daniel Rueckert

Abstract: In the recent years, convolutional neural networks have transformed the field of medical image analysis due to their capacity to learn discriminative image features for a variety of classification and regression tasks. However, successfully learning these features requires a large amount of manually annotated data, which is expensive to acquire and limited by the available resources of expert imag… ▽ More In the recent years, convolutional neural networks have transformed the field of medical image analysis due to their capacity to learn discriminative image features for a variety of classification and regression tasks. However, successfully learning these features requires a large amount of manually annotated data, which is expensive to acquire and limited by the available resources of expert image analysts. Therefore, unsupervised, weakly-supervised and self-supervised feature learning techniques receive a lot of attention, which aim to utilise the vast amount of available data, while at the same time avoid or substantially reduce the effort of manual annotation. In this paper, we propose a novel way for training a cardiac MR image segmentation network, in which features are learnt in a self-supervised manner by predicting anatomical positions. The anatomical positions serve as a supervisory signal and do not require extra manual annotation. We demonstrate that this seemingly simple task provides a strong signal for feature learning and with self-supervised learning, we achieve a high segmentation accuracy that is better than or comparable to a U-net trained from scratch, especially at a small data setting. When only five annotated subjects are available, the proposed method improves the mean Dice metric from 0.811 to 0.852 for short-axis image segmentation, compared to the baseline U-net. △ Less

Submitted 5 July, 2019; originally announced July 2019.

Comments: Accepted by MICCAI 2019

3D Cardiac Shape Prediction with Deep Neural Networks: Simultaneous Use of Images and Patient Metadata

Authors: Rahman Attar, Marco Pereanez, Christopher Bowles, Stefan K. Piechnik, Stefan Neubauer, Steffen E. Petersen, Alejandro F. Frangi

Abstract: Large prospective epidemiological studies acquire cardiovascular magnetic resonance (CMR) images for pre-symptomatic populations and follow these over time. To support this approach, fully automatic large-scale 3D analysis is essential. In this work, we propose a novel deep neural network using both CMR images and patient metadata to directly predict cardiac shape parameters. The proposed method u… ▽ More Large prospective epidemiological studies acquire cardiovascular magnetic resonance (CMR) images for pre-symptomatic populations and follow these over time. To support this approach, fully automatic large-scale 3D analysis is essential. In this work, we propose a novel deep neural network using both CMR images and patient metadata to directly predict cardiac shape parameters. The proposed method uses the promising ability of statistical shape models to simplify shape complexity and variability together with the advantages of convolutional neural networks for the extraction of solid visual features. To the best of our knowledge, this is the first work that uses such an approach for 3D cardiac shape prediction. We validated our proposed CMR analytics method against a reference cohort containing 500 3D shapes of the cardiac ventricles. Our results show broadly significant agreement with the reference shapes in terms of the estimated volume of the cardiac ventricles, myocardial mass, 3D Dice, and mean and Hausdorff distance. △ Less

Submitted 2 July, 2019; originally announced July 2019.

Comments: Accepted for publication in MICCAI 2019

Improving the generalizability of convolutional neural network-based segmentation on CMR images

Authors: Chen Chen, Wenjia Bai, Rhodri H. Davies, Anish N. Bhuva, Charlotte Manisty, James C. Moon, Nay Aung, Aaron M. Lee, Mihir M. Sanghvi, Kenneth Fung, Jose Miguel Paiva, Steffen E. Petersen, Elena Lukaschuk, Stefan K. Piechnik, Stefan Neubauer, Daniel Rueckert

Abstract: Convolutional neural network (CNN) based segmentation methods provide an efficient and automated way for clinicians to assess the structure and function of the heart in cardiac MR images. While CNNs can generally perform the segmentation tasks with high accuracy when training and test images come from the same domain (e.g. same scanner or site), their performance often degrades dramatically on ima… ▽ More Convolutional neural network (CNN) based segmentation methods provide an efficient and automated way for clinicians to assess the structure and function of the heart in cardiac MR images. While CNNs can generally perform the segmentation tasks with high accuracy when training and test images come from the same domain (e.g. same scanner or site), their performance often degrades dramatically on images from different scanners or clinical sites. We propose a simple yet effective way for improving the network generalization ability by carefully designing data normalization and augmentation strategies to accommodate common scenarios in multi-site, multi-scanner clinical imaging data sets. We demonstrate that a neural network trained on a single-site single-scanner dataset from the UK Biobank can be successfully applied to segmenting cardiac MR images across different sites and different scanners without substantial loss of accuracy. Specifically, the method was trained on a large set of 3,975 subjects from the UK Biobank. It was then directly tested on 600 different subjects from the UK Biobank for intra-domain testing and two other sets for cross-domain testing: the ACDC dataset (100 subjects, 1 site, 2 scanners) and the BSCMR-AS dataset (599 subjects, 6 sites, 9 scanners). The proposed method produces promising segmentation results on the UK Biobank test set which are comparable to previously reported values in the literature, while also performing well on cross-domain test sets, achieving a mean Dice metric of 0.90 for the left ventricle, 0.81 for the myocardium and 0.82 for the right ventricle on the ACDC dataset; and 0.89 for the left ventricle, 0.83 for the myocardium on the BSCMR-AS dataset. The proposed method offers a potential solution to improve CNN-based model generalizability for the cross-scanner and cross-site cardiac MR image segmentation task. △ Less

Submitted 3 July, 2019; v1 submitted 2 July, 2019; originally announced July 2019.

Comments: 15 pages, 8 figures

Using Harmonic Decomposition to Understand Barred Galaxy Evolution

Authors: Michael S. Petersen, Martin D. Weinberg, Neal Katz

Abstract: We study the mechanisms and evolutionary phases of bar formation in n-body simulations of a stellar disc and dark matter halo system using harmonic basis function expansion analysis to characterize the dynamical mechanisms in bar evolution. We correlate orbit families with phases of bar evolution by using empirical orthogonal functions that act as a spatial filter and form the gravitational potent… ▽ More We study the mechanisms and evolutionary phases of bar formation in n-body simulations of a stellar disc and dark matter halo system using harmonic basis function expansion analysis to characterize the dynamical mechanisms in bar evolution. We correlate orbit families with phases of bar evolution by using empirical orthogonal functions that act as a spatial filter and form the gravitational potential basis. In both models we find evidence for three phases in evolution with unique harmonic signatures. We recover known analytic results, such as bar slowdown owing to angular momentum transfer. We also find new dynamical mechanisms for bar evolution: a steady-state equilibrium configuration and harmonic interaction resulting in harmonic mode locking, both of which may be observable. Additionally, we find that ellipse fitting may severely overestimate measurements of bar length by a factor of two relative to the measurements based on orbits that comprise the true backbone supporting the bar feature. The bias will lead to overestimates of both bar mass and bar pattern speed, affecting inferences about the evolution of bars in the real universe, such as the fraction of bars with fast pattern speeds. We propose a direct observational technique to compute the radial extent of trapped orbits and determine a dynamical length for the bar. △ Less

Submitted 19 March, 2019; originally announced March 2019.

Comments: Submitted to MNRAS, comments welcome

Using Torque to Understand Barred Galaxy Models

Authors: Michael S. Petersen, Martin D. Weinberg, Neal Katz

Abstract: We track the angular momentum transfer in n-body simulations of barred galaxies by measuring torques to understand the dynamical mechanisms responsible for the evolution of the bar-disc-dark matter halo system. We find evidence for three distinct phases of barred galaxy evolution: assembly, secular growth, and steady-state equilibrium. Using a decomposition of the disc into orbital families, we tr… ▽ More We track the angular momentum transfer in n-body simulations of barred galaxies by measuring torques to understand the dynamical mechanisms responsible for the evolution of the bar-disc-dark matter halo system. We find evidence for three distinct phases of barred galaxy evolution: assembly, secular growth, and steady-state equilibrium. Using a decomposition of the disc into orbital families, we track bar mass and angular momentum through time and correlate the quantities with the phases of evolution. We follow the angular momentum transfer between particles and identify the dominant torque channels. We find that the halo model mediates the assembly and growth of the bar for a high central density halo, and the outer disc mediates the assembly and growth of the bar in a low central density halo model. Both galaxies exhibit a steady-state equilibrium phase where the bar is neither lengthening nor slowing. The steady-state equilibrium results from the balance of torque between particles that are gaining and losing angular momentum. We propose observational metrics for barred galaxies that can be used to help determine the evolutionary phase of a barred galaxy, and discuss the implications of the phases for galaxy evolution as a whole. △ Less

Submitted 8 October, 2019; v1 submitted 6 March, 2019; originally announced March 2019.

Comments: Accepted for publication in MNRAS

Unsupervised shape and motion analysis of 3822 cardiac 4D MRIs of UK Biobank

Authors: Qiao Zheng, Hervé Delingette, Kenneth Fung, Steffen E. Petersen, Nicholas Ayache

Abstract: We perform unsupervised analysis of image-derived shape and motion features extracted from 3822 cardiac 4D MRIs of the UK Biobank. First, with a feature extraction method previously published based on deep learning models, we extract from each case 9 feature values characterizing both the cardiac shape and motion. Second, a feature selection is performed to remove highly correlated feature pairs.… ▽ More We perform unsupervised analysis of image-derived shape and motion features extracted from 3822 cardiac 4D MRIs of the UK Biobank. First, with a feature extraction method previously published based on deep learning models, we extract from each case 9 feature values characterizing both the cardiac shape and motion. Second, a feature selection is performed to remove highly correlated feature pairs. Third, clustering is carried out using a Gaussian mixture model on the selected features. After analysis, we identify two small clusters which probably correspond to two pathological categories. Further confirmation using a trained classification model and dimensionality reduction tools is carried out to support this discovery. Moreover, we examine the differences between the other large clusters and compare our measures with the ground-truth. △ Less

Submitted 15 February, 2019; originally announced February 2019.

Using Commensurabilities and Orbit Structure to Understand Barred Galaxy Evolution

Authors: Michael S. Petersen, Martin D. Weinberg, Neal Katz

Abstract: We interpret simulations of secularly-evolving disc galaxies through orbit morphology. Using a new algorithm that measures the volume of orbits in real space using a tessellation, we rapidly isolate commensurate (resonant) orbits. We identify phase-space regions occupied by different orbital families. Compared to spectral methods, the tessellation algorithm can identify resonant orbits within a fe… ▽ More We interpret simulations of secularly-evolving disc galaxies through orbit morphology. Using a new algorithm that measures the volume of orbits in real space using a tessellation, we rapidly isolate commensurate (resonant) orbits. We identify phase-space regions occupied by different orbital families. Compared to spectral methods, the tessellation algorithm can identify resonant orbits within a few dynamical periods, crucial for understanding an evolving galaxy model. The flexible methodology accepts arbitrary potentials, enabling detailed descriptions of the orbital families. We apply the machinery to four different potential models, including two barred models, and fully characterise the orbital membership. We identify key differences in the content of orbit families, emphasising the presence of orbit families indicative of the bar evolutionary state and the shape of the dark matter halo. We use the characterisation of orbits to investigate the shortcomings of analytic and self-consistent studies, comparing our findings to the evolutionary epochs in self-consistent barred galaxy simulations. Using insight from our orbit analysis, we present a new observational metric that uses spatial and kinematic information from integral field spectrometers that may reveal signatures of commensurabilities and allow for a differentiation between models. △ Less

Submitted 29 April, 2021; v1 submitted 13 February, 2019; originally announced February 2019.

Automated Quality Control in Image Segmentation: Application to the UK Biobank Cardiac MR Imaging Study

Authors: Robert Robinson, Vanya V. Valindria, Wenjia Bai, Ozan Oktay, Bernhard Kainz, Hideaki Suzuki, Mihir M. Sanghvi, Nay Aung, Jos$é$ Miguel Paiva, Filip Zemrak, Kenneth Fung, Elena Lukaschuk, Aaron M. Lee, Valentina Carapella, Young Jin Kim, Stefan K. Piechnik, Stefan Neubauer, Steffen E. Petersen, Chris Page, Paul M. Matthews, Daniel Rueckert, Ben Glocker

Abstract: Background: The trend towards large-scale studies including population imaging poses new challenges in terms of quality control (QC). This is a particular issue when automatic processing tools, e.g. image segmentation methods, are employed to derive quantitative measures or biomarkers for later analyses. Manual inspection and visual QC of each segmentation isn't feasible at large scale. However, i… ▽ More Background: The trend towards large-scale studies including population imaging poses new challenges in terms of quality control (QC). This is a particular issue when automatic processing tools, e.g. image segmentation methods, are employed to derive quantitative measures or biomarkers for later analyses. Manual inspection and visual QC of each segmentation isn't feasible at large scale. However, it's important to be able to automatically detect when a segmentation method fails so as to avoid inclusion of wrong measurements into subsequent analyses which could lead to incorrect conclusions. Methods: To overcome this challenge, we explore an approach for predicting segmentation quality based on Reverse Classification Accuracy, which enables us to discriminate between successful and failed segmentations on a per-cases basis. We validate this approach on a new, large-scale manually-annotated set of 4,800 cardiac magnetic resonance scans. We then apply our method to a large cohort of 7,250 cardiac MRI on which we have performed manual QC. Results: We report results used for predicting segmentation quality metrics including Dice Similarity Coefficient (DSC) and surface-distance measures. As initial validation, we present data for 400 scans demonstrating 99% accuracy for classifying low and high quality segmentations using predicted DSC scores. As further validation we show high correlation between real and predicted scores and 95% classification accuracy on 4,800 scans for which manual segmentations were available. We mimic real-world application of the method on 7,250 cardiac MRI where we show good agreement between predicted quality metrics and manual visual QC scores. Conclusions: We show that RCA has the potential for accurate and fully automatic segmentation QC on a per-case basis in the context of large-scale population imaging as in the UK Biobank Imaging Study. △ Less

Submitted 27 January, 2019; originally announced January 2019.

Comments: 14 pages, 7 figures, Journal of Cardiovascular Magnetic Resonance