A5011338503- Advanced MRI Techniques and Applications
- Cardiovascular Function and Risk Factors
- Cardiac Imaging and Diagnostics
- Fetal and Pediatric Neurological Disorders
- Medical Image Segmentation Techniques
- Craniofacial Disorders and Treatments
- Cleft Lip and Palate Research
- Cardiac Valve Diseases and Treatments
- Medical Imaging Techniques and Applications
- Neonatal and fetal brain pathology
- Advanced Image Fusion Techniques
- Advanced Neural Network Applications
- Image and Signal Denoising Methods
- Fluid Dynamics Simulations and Interactions
- Underwater Acoustics Research
- Neurological disorders and treatments
- Advanced X-ray and CT Imaging
- Cardiac pacing and defibrillation studies
- Ultrasound Imaging and Elastography
- Elasticity and Material Modeling
- Blind Source Separation Techniques
- Remote-Sensing Image Classification
- Atomic and Subatomic Physics Research
- Radiomics and Machine Learning in Medical Imaging
- Ultrasonics and Acoustic Wave Propagation
University of Virginia
2022-2024
University of Virginia Health System
2021-2023
Centre National de la Recherche Scientifique
2023
Inserm
2017-2023
Institut National des Sciences Appliquées de Lyon
2023
University of Missouri
2023
Université de Picardie Jules Verne
2008-2020
K.N.Toosi University of Technology
2007-2020
Groupe de Recherches sur l’Analyse Multimodale de la Fonction Cérébrale
2020
To develop a three-dimensional (two dimensions + time) convolutional neural network trained with displacement encoding stimulated echoes (DENSE) data for and strain analysis of cine MRI.In this retrospective multicenter study, deep learning model (StrainNet) was developed to predict intramyocardial from contour motion. Patients various heart diseases healthy controls underwent cardiac MRI examinations DENSE between August 2008 January 2022. Network training inputs were time series myocardial...
Cardiovascular magnetic resonance (CMR) cine displacement encoding with stimulated echoes (DENSE) measures heart motion by myocardial into the signal phase, facilitating high accuracy and reproducibility of global segmental strain providing benefits in clinical performance. While conventional methods for analysis DENSE images are faster than those tagging, they still require manual user assistance. The present study developed evaluated deep learning fully-automatic analysis.
While multiple cardiovascular magnetic resonance (CMR) methods provide excellent reproducibility of global circumferential and longitudinal strain, achieving highly reproducible segmental strain is more challenging. Previous single-center studies have demonstrated displacement encoding with stimulated echoes (DENSE) strain. The present study evaluated the DENSE for measurement whole-slice or (Ecc), (Ell) radial (Err) torsion, Ecc at centers. Six centers participated a total 81 subjects were...
This study presents a new approach for segmentation and reconstruction of newborn's skull including bones, fontanels, sutures from computed tomography (CT) images. The relies on propagation pair interacting smooth surfaces based geodesic active regions. These evolve in opposite directions; the exterior surface moves inward while interior one outward direction. moving are forced to stop when arriving at outer or inner cranial bones using edge information. Since fontanels not directly...
A fully automated method for segmentation of neonatal skull in Magnetic Resonance (MR) images source localization electrical/magnetic encephalography (EEG/MEG) signals is proposed. Finding the these shows origin an abnormality. We propose a hybrid algorithm which Bayesian classifying framework combined with Hopfield Neural Network (HNN) segmentation. Due to non-homogeneity intensities MR images, local statistical parameters are used adaptive training neural network based on classifier error....
Inappropriate results may be produced if one uses adult or pediatric atlases for evaluation of neonatal cerebral images morphological studies. This is mainly due to anatomical particularities typical this early stage development. In paper, we describe the construction a digital brain atlas from set neonates aged between 39 and 42 weeks. It consists probabilistic models brain, cerebrospinal fluid (CSF) skull. first step, selected are segmented automatically followed by manual correction....
In this paper, we present a novel automatic algorithm for scalp and skull segmentation in T1-weighted neonatal head MR images. First, the probabilistic atlases are constructed. Second, outer surface is extracted based on an active mesh method. Third, maximum number of boundary points corresponding to inner using constructed atlas set knowledge rules. next step, priori information anatomy atlas. Finally, fast sweeping, tagging level methods applied reconstruct surfaces from detected...
Neonatal MR templates are appropriate for brain structural analysis and spatial normalization. However, they do not provide the essential accurate details of cranial bones fontanels-sutures. Distinctly, CT images best contrast bone definition In this paper, we present, first time, an approach to create a fully registered bimodal MR-CT head template neonates with gestational age 39 42 weeks. Such is functional studies, which require precise geometry including Due special characteristics...
Recently, the use of wavelet transform has led to significant advances in image denoising applications. Among wavelet-based approaches, Bayesian techniques give more accurate estimates. Considering interscale dependencies, these estimates become closer original image. In this context, choice an appropriate model for coefficients is important issue. The performance can also be improved by estimating parameters a local neighborhood. paper, we propose bivariate normal inverse Gaussian (NIG)...
This paper presents a novel method to automatically identify late-activating regions of the left ventricle from cine Displacement Encoding with Stimulated Echo (DENSE) MR images. We develop deep learning framework that identifies late mechanical activation in heart failure patients by detecting Time Onset circumferential Shortening (TOS). In particular, we build cascade network performing end-to-end (i) segmentation analyze cardiac function, (ii) prediction TOS based on spatiotemporal...
Introduction In displacement encoding with stimulated echoes (DENSE), tissue is encoded in the signal phase such that of each pixel space and time provides an independent measurement absolute displacement. Previously for DENSE, estimation Lagrangian used two steps: first a spatial interpolation and, second, least squares fitting through to Fourier or polynomial model. However, there no strong rationale through-time model, Methods To compute field from DENSE data, minimization problem...
Motivation: The need to address the challenge of a high nonresponse rate (approximately 40%) in CRT patients. Goal(s): By leveraging advanced computational methods, this research seeks redefine risk stratification and long-term survival prediction. Approach: 3D-CAE model is designed compress displacement trajectories into low-dimensional latent code while preserving sufficient information for trajectory reconstruction. network utilizes features from three specific slices predicting 4-year...
Recently, the use of wavelet transform has led to significant advances in image denoising applications. Among based approaches, Bayesian techniques give more accurate estimates. Considering interscale dependencies, these estimates become closer original image. In this context, choice an appropriate model for coefficients is important issue. The performance can also be improved by estimating parameters a local neighborhood. paper, we introduce spatially adaptive MMSE-based estimator using...
Cine DENSE provides both myocardial contours and intramyocardial displacements. We propose to use train deep networks predict motion from contour motion. Two workflows were implemented: a two-step FlowNet2-based framework with through-time correction network 3D (2D+t) Unet framework. Both depicted cardiac contraction abnormal patterns. The showed excellent reliability for global circumferential strain (E cc ) good segmental E , it outperformed commercial FT .