Article
Mathematics, Applied
Huan Han, Qi Jin, Yimin Zhang
Summary: In this paper, the authors relax the equality constraints of diffeomorphic image registration from Han and Wang (2020) [9] to include inequality constraints. Additionally, the issue of mesh folding is addressed, and a numerical algorithm for the proposed model is introduced. Numerical tests demonstrate the competitiveness of the proposed algorithm compared to other published image registration models.
COMPUTERS & MATHEMATICS WITH APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Jiulong Liu, Angelica I. Aviles-Rivero, Hui Ji, Carola-Bibiane Schonlieb
Summary: The paper introduces a novel hybrid method that intertwines indirect registration and reconstruction tasks in a single functional, aiming to reconstruct high quality images from few measurements with low computational cost. Through extensive experiments, the framework outperforms classic reconstruction schemes and other bi-task methods in terms of both image quality and computational time, demonstrating generalization capabilities on various imaging tasks.
MEDICAL IMAGE ANALYSIS
(2021)
Article
Computer Science, Artificial Intelligence
Huabing Zhou, Zhichao Xu, Yulu Tian, Zhenghong Yu, Yanduo Zhang, Jiayi Ma
Summary: This paper addresses image/surface deformation problem using semi-supervised learning, estimating interpolation functions through least squares and manifold regularization, and reducing computational complexity through sparse approximation. Experimental results show that the proposed method outperforms existing ones in image deformation.
PATTERN RECOGNITION
(2022)
Article
Biochemical Research Methods
Ameneh Sheikhjafari, Deepa Krishnaswamy, Michelle Noga, Nilanjan Ray, Kumaradevan Punithakumar
Summary: This study proposes an end-to-end supervised cardiac MRI segmentation framework based on diffeomorphic deformable registration for automatic segmentation of cardiac chambers. The method parameterizes the transformation using deep learning to represent cardiac deformation and ensures invertible transformations and prevents mesh folding for preserving the topology of the segmentation results. Experimental results show significant improvements of the proposed method compared to traditional methods in terms of Dice score and Hausdorff distance metrics.
IEEE TRANSACTIONS ON NANOBIOSCIENCE
(2023)
Article
Computer Science, Artificial Intelligence
Xiaogang Du, Dongxin Gu, Tao Lei, Song Wang, Xuejun Zhang, Hongying Meng
Summary: The proposed Hierarchical Neighborhood Spectral Features Log-Demons registration algorithm improves the capability of capturing complex distortions compared to traditional algorithms, and achieves better registration accuracy and efficiency.
IET IMAGE PROCESSING
(2021)
Article
Computer Science, Artificial Intelligence
Jianping Zhang, Yanyan Li
Summary: The paper proposes a new registration model based on an optimal control relaxation constraint for large deformation images, which guarantees that the registration mapping is diffeomorphic. An analysis of optimal control relaxation for indirectly seeking the diffeomorphic transformation is presented, along with a fast iterative scheme to solve the control increment optimization problem. The proposed model achieves state-of-the-art performance in quantitative evaluations and guarantees a diffeomorphic solution in the presence of large deformations.
SIAM JOURNAL ON IMAGING SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Ameneh Sheikhjafari, Michelle Noga, Kumaradevan Punithakumar, Nilanjan Ray
Summary: In this paper, a novel training-free approach based on ordinary differential equations is proposed for diffeomorphic deformable image registration. The approach employs an Euler integration type recursive scheme to estimate spatial transformations between fixed and moving images at different resolutions. Unlike learning-based methods, this approach does not require a dedicated training set and is not affected by training bias.
APPLIED INTELLIGENCE
(2022)
Article
Neurosciences
Jingru Fu, Antonios Tzortzakakis, Jose Barroso, Eric Westman, Daniel Ferreira, Rodrigo Moreno
Summary: Predicting brain aging is important for early detection and prognosis of neurodegenerative diseases. This paper proposes a methodology to fill missing data in longitudinal cohorts with anatomically plausible images. The proposed methodology uses deep learning-based diffeomorphic registration to simulate the aging process and rearrange the generated images to specific age ranges. The experimental results show that this methodology can produce anatomically plausible aging predictions and enhance longitudinal datasets.
HUMAN BRAIN MAPPING
(2023)
Article
Mathematics
Chenwei Cai, Lvda Wang, Shihui Ying
Summary: Image registration, an important technique in brain imaging analysis, aims to align two images through a spatial transformation. This research proposes a symmetric diffeomorphic image registration model based on multi-label segmentation masks. By introducing a new similarity metric and adaptive parameters, the proposed model improves accuracy, robustness, and smoothness of the registration compared to mainstream methods.
Article
Engineering, Biomedical
Nima Masoumi, Hassan Rivaz, M. Omair Ahmad, Yiming Xiao
Summary: The study proposes a novel inter-modal/contrast diffeomorphic registration algorithm that combines the Robust PaTch-based cOrrelation Ratio metric and the Fourier-Approximated Lie Algebras (FLASH) algorithm for fast and accurate image alignment. The results demonstrate that the proposed algorithm achieves comparable or even better registration accuracy and produces smoother deformations compared to existing techniques.
INTERNATIONAL JOURNAL OF COMPUTER ASSISTED RADIOLOGY AND SURGERY
(2023)
Article
Mathematics, Applied
Huan Han, Andong Wang
Summary: The fast multi grid algorithm proposed in this paper provides a satisfactory image registration result in a shorter time with no mesh folding, demonstrating competitiveness compared to other algorithms.
JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS
(2021)
Article
Engineering, Electrical & Electronic
Yong Lee, Shuang Mei
Summary: This paper proposes a novel particle image velocimetry (PIV) technique called diffeomorphic PIV, which introduces a deformation field to describe particle displacement. Compared to traditional iterative PIV, diffeomorphic PIV warps images with a deformation vector field instead of a velocity field, leading to significant accuracy improvement.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2022)
Article
Computer Science, Artificial Intelligence
Alexander Mangulad Christgau, Alexis Arnaudon, Stefan Sommer
Summary: Models of stochastic image deformation are used to study the continuous-time stochastic effects on images caused by deforming the image domain. These models have applications in longitudinal medical image analysis, which involves both population trends and random subject-specific variation. In this study, a stochastic extension of the LDDMM models with evolutions governed by a stochastic EPDiff equation is considered. Moment approximations of the corresponding Ito diffusion are used to construct estimators for statistical inference in the full stochastic model. It is demonstrated that this approach, when implemented efficiently with automatic differentiation tools, can accurately estimate parameters encoding the spatial correlation of noise fields in images.
JOURNAL OF MATHEMATICAL IMAGING AND VISION
(2023)
Article
Mathematics, Interdisciplinary Applications
Huan Han, Zhengping Wang, Yimin Zhang
Summary: In this work, a hierarchical image registration model based on LDDMM framework is introduced. Unlike most published models, this approach achieves a smooth minimizer for the cost functional without regularization. The existence and convergence of solutions for the multiscale approach are proved and numerical tests show satisfactory results.
MULTISCALE MODELING & SIMULATION
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Michael Miller, Jean Fan, Daniel J. Tward
Summary: Spatially resolved transcriptomic imaging is a promising new technology that quantifies gene expression at different scales. New mathematical and computational tools are needed to compare similarities between images. By using multi scale diffeomorphic metric mapping on MERFISH images, optimal nonrigid alignments between samples can be achieved.
2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS, CVPRW 2021
(2021)
Article
Neurosciences
Yassine Taoudi-Benchekroun, Daan Christiaens, Irina Grigorescu, Oliver Gale-Grant, Andreas Schuh, Maximilian Pietsch, Andrew Chew, Nicholas Harper, Shona Falconer, Tanya Poppe, Emer Hughes, Jana Hutter, Anthony N. Price, J-Donald Tournier, Lucilio Cordero-Grande, Serena J. Counsell, Daniel Rueckert, Tomoki Arichi, Joseph Hajnal, A. David Edwards, Maria Deprez, Dafnis Batalle
Summary: This study used machine learning methods to analyze structural connectomes of neonates and predict demographic and neurodevelopmental characteristics. The results demonstrate that a neural substrate of brain maturation with implications for future neurodevelopment can be detected from the neonatal connectome.
Article
Engineering, Electrical & Electronic
Kerstin Hammernik, Thomas Kustner, Burhaneddin Yaman, Zhengnan Huang, Daniel Rueckert, Florian Knoll, Mehmet Akcakaya
Summary: Physics-driven deep learning methods have revolutionized computational MRI reconstruction by improving the performance of reconstruction. This article provides an overview of recent developments in incorporating physics information into learning-based MRI reconstruction. It discusses both linear and non-linear forward models for computational MRI, classical approaches for solving these inverse problems, as well as physics-driven deep learning approaches such as physics-driven loss functions, plug-and-play methods, generative models, and unrolled networks. Challenges specific to MRI with linear and non-linear forward models are highlighted, and common issues and open challenges are also discussed.
IEEE SIGNAL PROCESSING MAGAZINE
(2023)
Article
Computer Science, Interdisciplinary Applications
Qiang Ma, Liu Li, Emma C. Robinson, Bernhard Kainz, Daniel Rueckert, Amir Alansary
Summary: CortexODE is a deep learning framework that uses neural ordinary differential equations (ODEs) to reconstruct cortical surfaces. By modeling the trajectories of points on the surface as ODEs and parameterizing the derivatives with a learnable deformation network, CortexODE is able to prevent self-intersections. Integrated with an automatic learning-based pipeline, CortexODE can efficiently reconstruct cortical surfaces in less than 5 seconds.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2023)
Article
Computer Science, Artificial Intelligence
Chen Qin, Shuo Wang, Chen Chen, Wenjia Bai, Daniel Rueckert
Summary: This paper proposes a novel method for myocardial motion tracking by using a generative model based on variational autoencoder to learn biomechanically plausible deformations and embed them into a neural network-parameterized transformation model. Experimental results show that the proposed method outperforms other approaches in terms of motion tracking accuracy, volume preservation, and generalizability.
MEDICAL IMAGE ANALYSIS
(2023)
Review
Mathematics, Applied
Darryl D. Holm, Wei Pan
Summary: This paper derives three different types of stochastic parameterisations for the interactions among disparate scales of motion in fluid convection. These models can be used to estimate prediction uncertainty.
PHYSICA D-NONLINEAR PHENOMENA
(2023)
Article
Computer Science, Interdisciplinary Applications
Cheng Ouyang, Chen Chen, Surui Li, Zeju Li, Chen Qin, Wenjia Bai, Daniel Rueckert
Summary: In this work, the authors investigate the problem of training a deep network that is robust to unseen domains using only data from one source domain. They propose a causality-inspired data augmentation approach to expose the model to synthesized domain-shifted training examples. The approach is validated on three cross-domain segmentation scenarios and shows consistent performance improvements compared to competitive methods.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2023)
Article
Computer Science, Artificial Intelligence
Tamara T. Mueller, Johannes C. Paetzold, Chinmay Prabhakar, Dmitrii Usynin, Daniel Rueckert, Georgios Kaissis
Summary: Graph Neural Networks (GNNs) have become the state-of-the-art for many machine learning applications, but differentially private training of GNNs has remained under-explored. In this work, we propose a framework for differentially private graph-level classification using DP-SGD, which is applicable to multi-graph datasets.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Robert Wright, Alberto Gomez, Veronika A. Zimmer, Nicolas Toussaint, Bishesh Khanal, Jacqueline Matthew, Emily Skelton, Bernhard Kainz, Daniel Rueckert, Joseph V. Hajnal, Julia A. Schnabel
Summary: This paper introduces a novel method to fuse partially imaged fetal head anatomy from multiple views into a single coherent 3D volume. The method aligns and fuses ultrasound images to improve image detail and minimize artifacts, achieving state-of-the-art performance in terms of image quality and robustness.
MEDICAL IMAGE ANALYSIS
(2023)
Article
Psychiatry
Hai Le, Konstantina Dimitrakopoulou, Hamel Patel, Charles Curtis, Lucilio Cordero-Grande, A. David Edwards, Joseph Hajnal, Jacques-Donald Tournier, Maria Deprez, Harriet Cullen
Summary: Increasing evidence suggests that deviations from normal early development may contribute to the onset of schizophrenia in adolescence and young adulthood. This study examined brain imaging changes associated with schizophrenia variants in newborns. The results revealed negative associations between schizophrenia genetic risk scores and brain volumes in several regions, indicating possible involvement of schizophrenia risk genes in early brain growth.
TRANSLATIONAL PSYCHIATRY
(2023)
Article
Obstetrics & Gynecology
Carla L. Avena-Zampieri, Jana Hutter, Maria Deprez, Kelly Payette, Megan Hall, Alena Uus, Surabhi Nanda, Anna Milan, Paul T. Seed, Mary Rutherford, Anne Greenough, Lisa Story
Summary: This study used T2* relaxometry to assess the development of fetal lungs and found that T2* values increased with gestational age, possibly reflecting increased perfusion and metabolic requirements, as well as alterations in tissue composition. Evaluating the results in fetuses with pulmonary morbidity may improve prenatal prognostication and perinatal care planning.
AMERICAN JOURNAL OF OBSTETRICS & GYNECOLOGY MFM
(2023)
Article
Computer Science, Interdisciplinary Applications
Adam Marcus, Paul Bentley, Daniel Rueckert
Summary: The proposed study introduces a novel end-to-end multi-task transformer-based model for concurrent segmentation and age estimation of cerebral ischemic lesions. The method captures long-range dependencies using gated positional self-attention and CT-specific data augmentation, and can be effectively trained with low-data regimes in medical imaging. Experimental results demonstrate promising performance in lesion age classification, outperforming existing task-specific algorithms.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2023)
Article
Computer Science, Artificial Intelligence
Jiazhen Pan, Manal Hamdi, Wenqi Huang, Kerstin Hammernik, Thomas Kuestner, Daniel Rueckert
Summary: This article introduces a learning-based and unrolled MCMR framework that can achieve accurate and rapid CMR reconstruction, delivering artifacts-free motion estimation and high-quality reconstruction even at imaging acceleration rates up to 20x.
MEDICAL IMAGE ANALYSIS
(2024)
Proceedings Paper
Cardiac & Cardiovascular Systems
Michael Tanzer, Sea Hee Yook, Pedro Ferreira, Guang Yang, Daniel Rueckert, Sonia Nielles-Vallespin
Summary: This study compares the effects of different input types, dimensionalities, and input types on the performance of a deep learning-based model in accelerating cardiac DTI. The results show that simpler 2D real-valued models outperform 3D or complex models, and the best performance is achieved by a real-valued model trained using both magnitude and phase components.
STATISTICAL ATLASES AND COMPUTATIONAL MODELS OF THE HEART: REGULAR AND CMRXMOTION CHALLENGE PAPERS, STACOM 2022
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Philip Mueller, Georgios Kaissis, Congyu Zou, Daniel Rueckert
Summary: In this study, we propose a new text-supervised pre-training method called LoVT, which is specifically optimized for localized medical imaging tasks. Compared to other methods, LoVT performs better on localized tasks.
COMPUTER VISION, ECCV 2022, PT XXVI
(2022)
Proceedings Paper
Computer Science, Interdisciplinary Applications
Michael Tanzer, Pedro Ferreira, Andrew Scott, Zohya Khalique, Maria Dwornik, Dudley Pennell, Guang Yang, Daniel Rueckert, Sonia Nielles-Vallespin
Summary: Diffusion Tensor Cardiac Magnetic Resonance (DT-CMR) is a non-invasive technique that allows the investigation of microstructural arrangement of cardiomyocytes within the myocardium. The current inefficiency of DT-CMR restricts its clinical use. This study proposes a new approach based on Generative Adversarial Networks, Vision Transformers, and Ensemble Learning, which significantly improves the efficiency and maintains image quality.
MEDICAL IMAGE UNDERSTANDING AND ANALYSIS, MIUA 2022
(2022)