A5039127636- Advanced MRI Techniques and Applications
- Medical Imaging Techniques and Applications
- Knee injuries and reconstruction techniques
- Radiomics and Machine Learning in Medical Imaging
- Osteoarthritis Treatment and Mechanisms
- Total Knee Arthroplasty Outcomes
- Advanced Neuroimaging Techniques and Applications
- Lower Extremity Biomechanics and Pathologies
- Atomic and Subatomic Physics Research
- Sports injuries and prevention
- Advanced Neural Network Applications
- Advanced X-ray and CT Imaging
- Tendon Structure and Treatment
- Shoulder Injury and Treatment
- Advanced NMR Techniques and Applications
- MRI in cancer diagnosis
- COVID-19 diagnosis using AI
- Electron Spin Resonance Studies
- COVID-19 Clinical Research Studies
- Cardiac Imaging and Diagnostics
- Domain Adaptation and Few-Shot Learning
- Image and Signal Denoising Methods
- Medical Imaging and Analysis
- Medical Image Segmentation Techniques
- Hemostasis and retained surgical items
Harvard University
1990-2025
Athinoula A. Martinos Center for Biomedical Imaging
2022-2025
Massachusetts General Hospital
2009-2025
Xidian University
2024-2025
The University of Queensland
2014-2025
China-Japan Friendship Hospital
2022-2025
Tianjin Medical University General Hospital
2023-2025
Second Military Medical University
2024
Shanghai First People's Hospital
2024
Shanghai Jiao Tong University
2024
For diagnosis of coronavirus disease 2019 (COVID-19), a SARS-CoV-2 virus-specific reverse transcriptase polymerase chain reaction (RT-PCR) test is routinely used. However, this can take up to 2 d complete, serial testing may be required rule out the possibility false negative results and there currently shortage RT-PCR kits, underscoring urgent need for alternative methods rapid accurate patients with COVID-19. Chest computed tomography (CT) valuable component in evaluation suspected...
BackgroundThe ongoing outbreak of COVID-19 pneumonia is globally concerning. We aimed to investigate the clinical and CT features in pregnant women children with this disease, which have not been well reported.MethodsClinical data 59 patients from January 27 February 14, 2020 were retrospectively reviewed, including 14 laboratory-confirmed non-pregnant adults, 16 25 clinically-diagnosed women, 4 children. The analyzed compared.FindingsCompared adults group (n = 14), initial normal body...
Purpose To develop and evaluate the feasibility of deep learning approaches for magnetic resonance (MR) imaging–based attenuation correction (AC) (termed MRAC) in brain positron emission tomography (PET)/MR imaging. Materials Methods A PET/MR imaging AC pipeline was built by using a approach to generate pseudo computed tomographic (CT) scans from MR images. convolutional auto-encoder network trained identify air, bone, soft tissue volumetric head images coregistered CT data training. set 30...
Purpose To describe and evaluate a new fully automated musculoskeletal tissue segmentation method using deep convolutional neural network (CNN) three‐dimensional (3D) simplex deformable modeling to improve the accuracy efficiency of cartilage bone within knee joint. Methods A pipeline was built by combining semantic CNN 3D modeling. technique called SegNet applied as core perform high resolution pixel‐wise multi‐class classification. The refined output from preserve overall shape maintain...
Purpose To determine the feasibility of using a deep learning approach to detect cartilage lesions (including softening, fibrillation, fissuring, focal defects, diffuse thinning due degeneration, and acute injury) within knee joint on MR images. Materials Methods A fully automated learning-based lesion detection system was developed by segmentation classification convolutional neural networks (CNNs). Fat-suppressed T2-weighted fast spin-echo MRI data sets 175 patients with pain were...
Purpose To describe and evaluate a new segmentation method using deep convolutional neural network (CNN), 3D fully connected conditional random field (CRF), simplex deformable modeling to improve the efficiency accuracy of knee joint tissue segmentation. Methods A pipeline was built by combining semantic CNN, CRF, modeling. encoder‐decoder designed as core perform high resolution pixel‐wise multi‐class classification for 12 different structures. The CRF applied regularize contextual...
To investigate the feasibility of using a deep learning-based approach to detect an anterior cruciate ligament (ACL) tear within knee joint at MRI by arthroscopy as reference standard.A fully automated diagnosis system was developed two convolutional neural networks (CNNs) isolate ACL on MR images followed classification CNN structural abnormalities isolated ligament. With institutional review board approval, sagittal proton density-weighted and fat-suppressed T2-weighted fast spin-echo in...
To develop and evaluate the feasibility of a data-driven deep learning approach (deepAC) for positron-emission tomography (PET) image attenuation correction without anatomical imaging. A PET pipeline was developed utilizing to generate continuously valued pseudo-computed (CT) images from uncorrected 18F-fluorodeoxyglucose (18F-FDG) images. convolutional encoder-decoder network trained identify tissue contrast in volumetric co-registered CT data. set 100 retrospective 3D FDG head used train...
To develop and evaluate a novel deep learning-based image reconstruction approach called MANTIS (Model-Augmented Neural neTwork with Incoherent k-space Sampling) for efficient MR parameter mapping.MANTIS combines end-to-end convolutional neural network (CNN) mapping, incoherent undersampling, physical model as synergistic framework. The CNN mapping directly converts series of undersampled images straight into maps using supervised training. Signal fidelity is enforced by adding pathway...
We present MRiLab, a new comprehensive simulator for large-scale realistic MRI simulations on regular PC equipped with modern graphical processing unit (GPU). MRiLab combines tissue modeling numerical virtualization of an system and scanning experiment to enable assessment broad range approaches including advanced quantitative methods inferring microstructure sub-voxel level. A flexible representation is achieved in by employing the generalized model multiple exchanging water macromolecular...
To develop and evaluate a novel deep learning-based reconstruction framework called SANTIS (Sampling-Augmented Neural neTwork with Incoherent Structure) for efficient MR image improved robustness against sampling pattern discrepancy.With combination of data cycle-consistent adversarial network, end-to-end convolutional neural network mapping, fidelity enforcement reconstructing undersampled data, additionally utilizes sampling-augmented training strategy by extensively varying undersampling...
Abstract For diagnosis of COVID-19, a SARS-CoV-2 virus-specific reverse transcriptase polymerase chain reaction (RT-PCR) test is routinely used. However, this can take up to two days complete, serial testing may be required rule out the possibility false negative results, and there currently shortage RT-PCR kits, underscoring urgent need for alternative methods rapid accurate COVID-19 patients. Chest computed tomography (CT) valuable component in evaluation patients with suspected infection....
To develop a model-guided self-supervised deep learning MRI reconstruction framework called reference-free latent map extraction (RELAX) for rapid quantitative MR parameter mapping.Two physical models are incorporated network training in RELAX, including the inherent imaging model and that is used to fit parameters MRI. By enforcing these constraints, RELAX eliminates need full sampled reference data sets required standard supervised learning. Meanwhile, also enables direct of corresponding...
Abstract Purpose This paper proposes a novel self‐supervised learning framework that uses model reinforcement, REference‐free LAtent map eXtraction with MOdel REinforcement (RELAX‐MORE), for accelerated quantitative MRI (qMRI) reconstruction. The proposed method an optimization algorithm to unroll iterative model‐based qMRI reconstruction into deep framework, enabling MR parameter maps are highly accurate and robust. Methods Unlike conventional methods which require large amounts of training...
Retransmitted messages online can have profound effects on disaster response; however, existing literature provides an incomplete account of why are retransmitted social media in disasters. In particular, there is a need to theorize the capabilities communication tools used for sending messages, because nowadays people send via different tools. This paper aims and explain how affect message retransmission by affecting generation characteristics. To test our account, we collected coded...
Quantitative knowledge on the anatomy of medial collateral ligament (MCL) is important for treatment MCL injury and release during total knee arthroplasty (TKA). The objective this study was to quantitatively determine morphology human knees. 10 cadaveric knees were dissected investigate anatomy. specimens fixed in full extension position maintained dissection morphometric measurements. outlines insertion sites superficial (sMCL) deep (dMCL) digitized using a 3D digitizing system. areas...
Purpose To investigate the feasibility of using compressed sensing (CS) to accelerate three‐dimensional fast spin‐echo (3D‐FSE) imaging knee. Materials and Methods A 3D‐FSE sequence was performed at 3T with CS (CUBE‐CS 3:16‐minute scan time) without (CUBE 4:44‐minute twice on knees 10 healthy volunteers assess signal‐to‐noise ratio (SNR) addition‐subtraction method once 50 symptomatic patients diagnostic performance. SNR cartilage, muscle, synovial fluid, bone marrow CUBE CUBE‐CS images were...