Article
Biophysics
Amalia Villa, Sebastian Ingelaere, Ben Jacobs, Bert Vandenberk, Sabine Van Huffel, Rik Willems, Carolina Varon
Summary: This study proposes a unified framework using Laplacian Eigenmaps (LE) to compare different subjects' electrocardiographic (ECG) signals and enhance abnormal conditions. By mapping new subjects' signals onto a normal reference ECG space, ECG abnormalities can be captured and quantified. The results of the study demonstrate that this method can effectively detect pathological changes such as ischemic heart disease and dilated cardiomyopathy.
PHYSIOLOGICAL MEASUREMENT
(2023)
Article
Computer Science, Artificial Intelligence
Hongmin Cai, Xiaoqi Sheng, Guorong Wu, Bin Hu, Yiu-Ming Cheung, Jiazhou Chen
Summary: There is increasing evidence that Alzheimer's disease (AD) disrupts the brain network before clinical symptoms appear, allowing for early diagnosis. The current methods of analyzing brain networks treat the high-dimensional data as regular matrices or vectors, which leads to a loss of essential network topology and affects diagnosis accuracy. To address this issue, this article proposes a network manifold harmonic discriminant analysis (MHDA) method for accurately detecting AD. The effectiveness of the proposed method in stratifying cognitively normal controls, mild cognitive impairment, and AD is demonstrated through extensive experiments.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Chemistry, Analytical
Edward Harefa, Weidong Zhou
Summary: Laser-induced breakdown spectroscopy (LIBS) is a method for material classification, but extracting meaningful information from large amounts of spectral data and reducing dimensions is a challenge. This study used manifold dimensionality reduction techniques and a support vector machine model to classify five different types of aluminum alloys. The results showed that non-linear manifold learning techniques can achieve better classification performance.
Article
Multidisciplinary Sciences
Tristan Millington
Summary: When studying financial markets, we often estimate correlation matrices from asset returns. Due to noise and dimensionality, these matrices are filtered using methods like minimum spanning tree and graph theory. However, the assumption that the data fits a certain shape is not necessarily valid, and empirical investigations comparing these methods are few. This paper examines how filtered networks differ from original networks using stock returns from different markets and evaluates their impact on classification accuracy. The findings suggest that the relationship between the full and filtered networks depends on the data and market conditions, and increasing network size weakens this relationship. Additionally, the filtered networks do not improve classification accuracy compared to the full networks.
Article
Spectroscopy
Xiao-Wen Zhang, Zheng-Guang Chen, Feng Jiao
Summary: Laplacian Eigenmaps is a nonlinear dimensionality reduction algorithm based on graph theory. In this paper, an adaptive LE improved algorithm is proposed to optimize the weight calculation by considering adjacent sample points and multi-scale data, achieving better dimensionality reduction effect.
SPECTROCHIMICA ACTA PART A-MOLECULAR AND BIOMOLECULAR SPECTROSCOPY
(2022)
Article
Computer Science, Information Systems
Sichao Fu, Weifeng Liu, Kai Zhang, Yicong Zhou, Dapeng Tao
Summary: This paper introduces the graph p-Laplacian convolutional networks (GpLCN) to better extract sample features and improve classification performance by utilizing graph p-Laplacian matrix.
INFORMATION SCIENCES
(2021)
Article
Automation & Control Systems
Guolin Yu, Jun Ma, Chenzhen Xie
Summary: This paper proposes a Hessian scatter regularized twin support vector machine (HSR-TSVM) based on Laplacian regularization. HSR-TSVM can better maintain the local topology of the samples and improve classification performance by utilizing the structural information of samples. Furthermore, a least-squares version of HSR-TSVM called HSR-LSTSVM is proposed to improve computational efficiency.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Chemistry, Physical
Jakub Rydzewski
Summary: Constructing reduced representations of high-dimensional systems is a fundamental problem in physical chemistry. This study addresses the issue of selecting high-dimensional representations before dimensionality reduction using a method called the reweighted diffusion map. The method quantitatively selects high-dimensional representations by exploring the spectral decomposition of Markov transition matrices built from data obtained from atomistic simulations.
JOURNAL OF PHYSICAL CHEMISTRY LETTERS
(2023)
Article
Biochemical Research Methods
Yulian Ding, Xiujuan Lei, Bo Liao, Fang-Xiang Wu
Summary: In this study, a novel model MLRDFM is proposed to predict miRNA-disease associations, which improves the performance of DeepFM by considering relationships among items and using Laplacian regularization. Experimental results show that MLRDFM enhances performance and reduces overfitting, outperforming state-of-the-art models in miRNA-disease association prediction.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Computer Science, Artificial Intelligence
Daniel Floryan, Michael D. D. Graham
Summary: Dynamical models are crucial for understanding and predicting natural systems, but the choice of state variables is often redundant and obscures the underlying behavior of the system. This study combines manifold theory and neural networks to develop a method that learns a system's intrinsic state variables directly from time-series data, reducing the dimensionality of the data.
NATURE MACHINE INTELLIGENCE
(2022)
Article
Neurosciences
Javier Gonzalez-Castillo, Isabel S. S. Fernandez, Ka Chun Lam, Daniel A. A. Handwerker, Francisco Pereira, Peter A. A. Bandettini
Summary: Whole-brain functional connectivity (FC) measured with functional MRI (fMRI) evolves over time and can be explored using manifold learning techniques (MLTs) to generate low dimensional representations. This study investigates the use of MLTs, such as Laplacian Eigenmaps (LEs), T-distributed Stochastic Neighbor Embedding (T-SNE), and Uniform Manifold Approximation and Projection (UMAP), to capture neuro-biologically meaningful representations of time-varying FC (tvFC) data. The results show that tvFC data has an intrinsic dimension (ID) ranging from 4 to 26, with significant variation between resting and task states. However, the application of MLTs to unlabeled data such as resting-state remains challenging.
FRONTIERS IN HUMAN NEUROSCIENCE
(2023)
Article
Computer Science, Artificial Intelligence
Ziyi Su, Wang Wenbo, Weibin Zhang
Summary: This paper proposes a novel method, called Regularized Denoising Latent Subspace based Linear Regression (RDLSLR), for noisy image classification. The RDLSLR model divides the traditional subspace learning model into two steps, adding a denoising latent space to obtain clean data and using another transformation matrix to learn regression target. It further optimizes the data distribution using Laplacian Regularization and introduces the e-dragging technique for enhanced discriminative power. Experimental results demonstrate the effectiveness of the RDLSLR model in various recognition tasks.
PATTERN ANALYSIS AND APPLICATIONS
(2023)
Article
Chemistry, Multidisciplinary
Rizwan Khan, Zahid Hussain Qaisar, Atif Mehmood, Ghulam Ali, Tamim Alkhalifah, Fahad Alturise, Lingna Wang
Summary: Patients with Alzheimer's disease go through irreversible stages, and early detection is crucial for disease progression. Diagnostic techniques rely on MRI and high-dimensional 3D imaging data. Deep learning-based methods can help detect different stages of AD, but face challenges with 3D volumes. This study proposes a deep learning-based multiclass classification method for early diagnosis of Alzheimer's, achieving high accuracy and outperforming existing methods.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Artificial Intelligence
Yao Zhang, Yingcang Ma
Summary: This paper proposes a non-negative multi-label feature selection method with dynamic graph constraints to address the loss of label information. Experimental results demonstrate the effectiveness of the proposed method.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Medicine, General & Internal
Samra Shahzadi, Naveed Anwer Butt, Muhammad Usman Sana, Inaki Elio Pascual, Mercedes Briones Urbano, Isabel de la Torre Diez, Imran Ashraf
Summary: This study used independent component analysis to investigate the impact of Alzheimer's disease on different brain regions at different stages. Machine learning algorithms were used to categorize the stages of the disease. The study found that certain regions were impacted in all stages, and AdaBoost algorithm achieved excellent classification results.
Article
Neurosciences
Sunniva Fenn-Moltu, Sean P. Fitzgibbon, Judit Ciarrusta, Michael Eyre, Lucilio Cordero-Grande, Andrew Chew, Shona Falconer, Oliver Gale-Grant, Nicholas Harper, Ralica Dimitrova, Katy Vecchiato, Daphna Fenchel, Ayesha Javed, Megan Earl, Anthony N. Price, Emer Hughes, Eugene P. Duff, Jonathan O'Muircheartaigh, Chiara Nosarti, Tomoki Arichi, Daniel Rueckert, Serena Counsell, Joseph Hajnal, A. David Edwards, Grainne McAlonan, Dafnis Batalle
Summary: The formation of the functional connectome in early life is crucial for future learning and behavior. However, our understanding of how the functional organization of brain regions matures during the early postnatal period, especially in response to adverse neurodevelopmental outcomes like preterm birth, is limited. In this study involving 366 neonates, we found that functional centrality (weighted degree) increased with age in visual regions and decreased in motor and auditory regions in term-born infants. Preterm-born infants scanned at term equivalent age showed higher functional centrality in visual regions and lower measures in motor regions. Functional centrality did not predict neurodevelopmental outcomes at 18 months old.
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
Radiology, Nuclear Medicine & Medical Imaging
Ayse Sila Dokumaci, Fraser R. Aitken, Jan Sedlacik, Pip Bridgen, Raphael Tomi-Tricot, Ronald Mooiweer, Katy Vecchiato, Tom Wilkinson, Chiara Casella, Sharon Giles, Joseph Hajnal, Shaihan J. Malik, Jonathan O'Muircheartaigh, David W. Carmichael
Summary: In this study, an optimized MP2RAGE protocol at 7 Tesla was developed to provide T1-weighted uniform image and gray matter-dominant fluid and white matter suppression contrast images simultaneously in a clinically applicable acquisition time. The results showed that high-contrast images with excellent anatomical detail could be obtained using the optimized parameter set.
MAGNETIC RESONANCE IN MEDICINE
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Francesco Padormo, Paul Cawley, Louise Dillon, Emer Hughes, Jennifer Almalbis, Joanna Robinson, Alessandra Maggioni, Miguel De La Fuente Botella, Dan Cromb, Anthony Price, Lori Arlinghaus, John Pitts, Tianrui Luo, Dingtian Zhang, Sean C. L. Deoni, Steve Williams, Shaihan Malik, Jonathan O'Muircheartaigh, Serena J. Counsell, Mary Rutherford, Tomoki Arichi, A. David Edwards, Joseph V. Hajnal
Summary: This study utilizes ultralow-field MRI systems to measure T-1 values in neonates and finds that these values are shorter than those previously measured at standard clinical field strengths, but longer than those of adults at ultralow-field. T-1 values decrease with postmenstrual age, making them a potential biomarker for perinatal brain development.
MAGNETIC RESONANCE IN MEDICINE
(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)
Letter
Anesthesiology
Philippa Bridgen, Shaihan Malik, Thomas Wilkinson, John N. Cronin, Tahzeeb Bhagat, Nicholas Hart, Stuart Mc Corkell, Joanne Perkins, Shane Tibby, Sara Hanna, Richard Kirwan, Thomas Pauly, Arthur Weeks, Geoff Charles-Edwards, Francesco Padormo, David Stell, Kariem El-Boghdadly, Sebastien Ourselin, Sharon L. Giles, Anthony D. Edwards, Joseph V. Hajnal, Benjamin J. Blaise
BRITISH JOURNAL OF ANAESTHESIA
(2023)
Article
Neurosciences
Abi Fukami-Gartner, Ana A. Baburamani, Ralica Dimitrova, Prachi A. Patkee, Olatz Ojinaga-Alfageme, Alexandra F. Bonthrone, Daniel Cromb, Alena U. Uus, Serena J. Counsell, Joseph Hajnal, Jonathan O'Muircheartaigh, Mary A. Rutherford
Summary: Down syndrome (DS) is a common genetic cause of intellectual disability. In this study, researchers analyzed the brain volumes of neonates with DS using neuroimaging techniques. They found that the DS brain showed significant reductions in overall volume, cerebral white matter, and cerebellar volumes, as well as differences in relative lobar volumes. Furthermore, certain features such as enlarged deep gray matter volume and lateral ventricle enlargement were observed. Assessing phenotypic severity at the neonatal stage may help guide early interventions and improve neurodevelopmental outcomes in children with DS.
Article
Computer Science, Interdisciplinary Applications
Lucilio Cordero-Grande, Juan Enrique Ortuno-Fisac, Alejandra Aguado del Hoyo, Alena Uus, Maria Deprez, Andres Santos, Joseph V. Hajnal, Maria J. Ledesma-Carbayo
Summary: In this paper, a deep generative prior and a diffeomorphic volume to slice registration method are proposed for robust volumetric reconstructions. Experiments on 72 fetal datasets show that our method outperforms existing techniques in improving image quality and accurately predicting gestational age.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(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
Biology
Sian Wilson, Maximilian Pietsch, Lucilio Cordero-Grande, Daan Christiaens, Alena Uus, Vyacheslav R. Karolis, Vanessa Kyriakopoulou, Kathleen Colford, Anthony N. Price, Jana Hutter, Mary A. Rutherford, Emer J. Hughes, Serena J. Counsell, Jacques-Donald Tournier, Joseph Hajnal, A. David Edwards, Jonathan O'Muicheartaigh, Tomoki Arichi, Finnegan J. Calabro
Summary: In this study, high-resolution in utero diffusion magnetic resonance imaging was used to examine the development of thalamocortical white matter in 140 fetuses. The researchers delineated the thalamocortical pathways and parcellated the fetal thalamus based on its cortical connectivity. They quantified microstructural tissue components along the tracts in fetal compartments and identified changes in diffusion metrics reflecting critical neurobiological transitions. These findings provide a normative reference for further studies on developmental disruptions and their contributions to pathophysiology.
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
Automation & Control Systems
Xianqiang Bao, Shuangyi Wang, Lingling Zheng, Richard James Housden, Joseph Hajnal, Kawal Rhode
Summary: This article proposes a novel self-adaptive parallel manipulator (SAPM) for robotic ultrasonography. The SAPM can automatically adjust the ultrasound probe pose, provide approximate constant operating forces/torques, achieve mechanical measurement, and cushion undesired produced forces. Experimental results show that the SAPM can provide 3 DOFs motion, operating force/torque measurement, and automatically adjust the US probe pose to capture high-quality ultrasound images.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2023)
Article
Cardiac & Cardiovascular Systems
Daniel Cromb, Alexandra F. Bonthrone, Alessandra Maggioni, Paul Cawley, Ralica Dimitrova, Christopher J. Kelly, Lucilio Cordero-Grande, Olivia Carney, Alexia Egloff, Emer Hughes, Joseph V. Hajnal, John Simpson, Kuberan Pushparajah, Mary A. Rutherford, A. David Edwards, Jonathan O'Muircheartaigh, Serena J. Counsell
Summary: Infants with congenital heart disease (CHD) are at risk of impaired brain growth, especially in the immediate postoperative period. The duration of postoperative intensive care stay is associated with the degree of impaired brain growth. Clinical risk factors, such as higher preoperative creatinine levels and longer cardiopulmonary bypass duration, are also associated with impaired brain growth.
JOURNAL OF THE AMERICAN HEART ASSOCIATION
(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)
Article
Computer Science, Artificial Intelligence
Hong Liu, Dong Wei, Donghuan Lu, Xiaoying Tang, Liansheng Wang, Yefeng Zheng
Summary: This study proposes a framework based on hybrid 2D-3D convolutional neural networks for obtaining continuous 3D retinal layer surfaces from OCT volumes. The framework works well with both full and sparse annotations and utilizes alignment displacement vectors and layer segmentation to align the B-scans and segment the layers. Experimental results show that the framework outperforms state-of-the-art 2D deep learning methods in terms of layer segmentation accuracy and cross-B-scan 3D continuity.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Simon Oxenford, Ana Sofia Rios, Barbara Hollunder, Clemens Neudorfer, Alexandre Boutet, Gavin J. B. Elias, Jurgen Germann, Aaron Loh, Wissam Deeb, Bryan Salvato, Leonardo Almeida, Kelly D. Foote, Robert Amaral, Paul B. Rosenberg, David F. Tang-Wai, David A. Wolk, Anna D. Burke, Marwan N. Sabbagh, Stephen Salloway, M. Mallar Chakravarty, Gwenn S. Smith, Constantine G. Lyketsos, Michael S. Okun, William S., Zoltan Mari, Francisco A. Ponce, Andres Lozano, Wolf-Julian Neumann, Bassam Al-Fatly, Andreas Horn
Summary: Spatial normalization is a method to map subject brain images to an average template brain, allowing comparison of brain imaging results. We introduce a novel tool called WarpDrive, which enables manual refinements of image alignment after automated registration. The tool improves accuracy of data representation and aids in understanding patient outcomes.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Ricards Marcinkevics, Patricia Reis Wolfertstetter, Ugne Klimiene, Kieran Chin-Cheong, Alyssia Paschke, Julia Zerres, Markus Denzinger, David Niederberger, Sven Wellmann, Ece Ozkan, Christian Knorr, Julia E. Vogt
Summary: This study presents interpretable machine learning models for predicting the diagnosis, management, and severity of suspected appendicitis using ultrasound images. The proposed models utilize concept bottleneck models (CBM) that facilitate interpretation and intervention by clinicians, without compromising performance or requiring time-consuming image annotation.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Jian-Qing Zheng, Ziyang Wang, Baoru Huang, Ngee Han Lim, Bartlomiej W. Papiez
Summary: This article introduces a new method for medical image registration, which utilizes a separable motion backbone and a residual aligner module to better handle the discontinuous motion of multiple neighboring objects. The proposed method achieves excellent registration results on abdominal CT scans and lung CT scans.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Xiangqiong Wu, Guanghua Tan, Hongxia Luo, Zhilun Chen, Bin Pu, Shengli Li, Kenli Li
Summary: This study develops a user-friendly framework for the automated diagnosis of thyroid nodules in ultrasound videos, simulating the diagnostic workflow of radiologists. By interpreting image characteristics and modeling temporal contextual information, the efficiency and generalizability of the diagnosis can be improved.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Riddhish Bhalodia, Shireen Elhabian, Jadie Adams, Wenzheng Tao, Ladislav Kavan, Ross Whitaker
Summary: This paper introduces DeepSSM, a deep learning-based framework for image-to-shape modeling. By learning the functional mapping from images to low-dimensional shape descriptors, DeepSSM can directly infer statistical representation of anatomy from 3D images. Compared to traditional methods, DeepSSM eliminates the need for heavy manual preprocessing and segmentation, and significantly improves computational time.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Florentin Liebmann, Marco von Atzigen, Dominik Stutz, Julian Wolf, Lukas Zingg, Daniel Suter, Nicola A. Cavalcanti, Laura Leoty, Hooman Esfandiari, Jess G. Snedeker, Martin R. Oswald, Marc Pollefeys, Mazda Farshad, Philipp Furnstahl
Summary: This study presents a marker-less approach for automatic registration and real-time navigation of lumbar spinal fusion surgery using a deep neural network, avoiding radiation exposure and surgical errors. The method was validated on an ex-vivo surgery and a public dataset.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Piyush Tiwary, Kinjawl Bhattacharyya, A. P. Prathosh
Summary: Domain shift refers to the change of distributional characteristics between training and testing datasets, leading to performance drop. For medical image tasks, domain shift can be caused by changes in imaging modalities, devices, and staining mechanisms. Existing approaches based on generative models suffer from training difficulties and lack of diversity. In this paper, the authors propose the use of energy-based models (EBMs) for unpaired image-to-image translation in medical images. The proposed method, called Cycle Consistent Twin EBMs (CCT-EBM), employs a pair of EBMs in the latent space of an Auto-Encoder to ensure translation symmetry and coupling between domains.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Yutong Xie, Jianpeng Zhang, Lingqiao Liu, Hu Wang, Yiwen Ye, Johan Verjans, Yong Xia
Summary: This paper proposes a hybrid pre-training paradigm that combines self-supervised learning and supervised learning to improve the representation quality for medical image segmentation tasks. It introduces a reference task in self-supervised learning and optimizes the model using a gradient matching method. The experimental results demonstrate the effectiveness of this approach on multiple medical image segmentation benchmarks.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Youyi Song, Jing Zou, Kup-Sze Choi, Baiying Lei, Jing Qin
Summary: Cell classification is crucial for intelligent cervical cancer screening, but the variation in cells' appearance and shape poses challenges. A new learning algorithm, worse-case boosting, is proposed to improve classification accuracy for under-represented data. Experimental results demonstrate the effectiveness of this algorithm in two publicly available datasets, achieving a 4% improvement in accuracy.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Sangjoon Park, Eun Sun Lee, Kyung Sook Shin, Jeong Eun Lee, Jong Chul Ye
Summary: The increasing demand for AI systems to monitor human errors and abnormalities in healthcare presents challenges. This study presents a model called Medical X-VL, which is tailored for the medical domain and outperformed current state-of-the-art models in two medical image datasets. The model enables various zero-shot tasks for monitoring AI in the medical domain.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Anna Klimovskaia Susmelj, Berkan Lafci, Firat Ozdemir, Neda Davoudi, Xose Luis Dean-Ben, Fernando Perez-Cruz, Daniel Razansky
Summary: Optoacoustic imaging is a technique that uses optical excitation and ultrasound detection for biological tissue imaging. The quality of the images depends on the extent of tomographic coverage provided by the ultrasound detector arrays. However, full coverage is not always possible due to experimental constraints. The proposed signal domain adaptation network aims to reduce limited-view artifacts in the images.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Srijay Deshpande, Muhammad Dawood, Fayyaz Minhas, Nasir Rajpoot
Summary: In this work, a novel framework called SynCLay is proposed for automated synthesis of histology images based on user-defined cellular layouts. The framework can generate realistic and high-quality histology images with different cellular arrangements, which is helpful for studying the role of cells in the tumor microenvironment. The framework integrates a nuclear segmentation and classification model to refine nuclear structures and generate nuclear masks. Evaluation using quantitative metrics and feedback from pathologists shows that the synthetic images generated by SynCLay have high realism scores and can accurately differentiate between benign and malignant tumors.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Ahmed H. Shahin, An Zhao, Alexander C. Whitehead, Daniel C. Alexander, Joseph Jacob, David Barber
Summary: Survival analysis is a valuable tool in healthcare for predicting the time to specific events. This paper introduces CenTime, a novel approach that directly estimates the time to event. The method performs well with censored data and can be easily integrated with deep learning models. Compared to standard methods, CenTime offers superior performance in predicting event time while maintaining comparable ranking performance.
MEDICAL IMAGE ANALYSIS
(2024)
Article
Computer Science, Artificial Intelligence
Bingyuan Liu, Jose Dolz, Adrian Galdran, Riadh Kobbi, Ismail Ben Ayed
Summary: Most segmentation losses, such as CE and Dice, are variants of the Cross-Entropy or Dice losses. This work provides a theoretical analysis that shows a deeper connection between CE and Dice than previously thought. From a constrained-optimization perspective, both CE and Dice decompose into similar ground-truth matching terms and region-size penalty terms. The analysis uncovers hidden region-size biases: Dice has an intrinsic bias towards extremely imbalanced solutions, while CE implicitly encourages the ground-truth region proportions. Based on this analysis, a principled and simple solution is proposed to explicitly control the region-size bias.
MEDICAL IMAGE ANALYSIS
(2024)