A5079952004- Medical Imaging Techniques and Applications
- Advanced X-ray and CT Imaging
- Radiation Dose and Imaging
- Advanced MRI Techniques and Applications
- Flame retardant materials and properties
- Traumatic Brain Injury and Neurovascular Disturbances
- Cardiac Arrest and Resuscitation
- Radiomics and Machine Learning in Medical Imaging
- Anesthesia and Neurotoxicity Research
- Advanced Neuroimaging Techniques and Applications
- Synthesis and properties of polymers
- Vascular Malformations Diagnosis and Treatment
- Fire dynamics and safety research
- Photoacoustic and Ultrasonic Imaging
- Image and Signal Denoising Methods
- Intracerebral and Subarachnoid Hemorrhage Research
- Head and Neck Cancer Studies
- MRI in cancer diagnosis
- Organ Transplantation Techniques and Outcomes
- Respiratory Support and Mechanisms
- Advanced machining processes and optimization
- Industrial Technology and Control Systems
- AI in cancer detection
- Medical Research and Treatments
- COVID-19 diagnosis using AI
China Academy of Engineering Physics
2023-2024
Chongqing University
2019-2024
Rensselaer Polytechnic Institute
2020-2024
University of Chicago
2009-2023
Wuhan University of Technology
2023
Chongqing University of Technology
2019-2023
Imaging Center
2012-2022
Union Hospital
2005-2020
Huazhong University of Science and Technology
2005-2020
Sichuan University
2013-2020
Grading laryngeal squamous cell carcinoma (LSCC) based on histopathological images is a clinically significant yet challenging task. However, more low-effect background semantic information appeared in the feature maps, channels, and class activation which caused serious impact accuracy interpretability of LSCC grading. While traditional transformer block makes extensive use parameter attention, model overlearns information, resulting ineffectively reducing proportion semantics. Therefore,...
Potential risk of X-ray radiation from computed tomography (CT) has been a concern the public. However, simply decreasing dose will degrade quality CT images and compromise diagnostic performance. In this paper, we propose noise learning generative adversarial network coupling with least squares, structural similarity L1 losses for low-dose denoising. our method, distributed in input image is learned by generator then subtracted to generate final denoised version. The are penalized squares...
Tomographic image reconstruction with deep learning is an emerging field of applied artificial intelligence. Reducing radiation dose sparse views’ a significant task in cardiac imaging. Many efforts are contributing to sparse-view tomography imaging, but it still challenge for achieving good images from high level, such as 60 views. In this study, we proposed Deep Embedding-Attention-Refinement (DEAR) network fundamentally address challenge. DEAR consists three modules including embedding,...
Image denoising is a prerequisite for downstream tasks in many fields. Low-dose and photon-counting computed tomography (CT) can optimize diagnostic performance at minimized radiation dose. Supervised deep methods are popular but require paired clean or noisy samples that often unavailable practice. Limited by the independent noise assumption, current self-supervised cannot process correlated noises as CT images. Here we propose first-of-its-kind similarity-based approach, referred to...
The dynamics of fluid flow in normal pressure hydrocephalus (NPH) are poorly understood. Normally, CSF flows out the brain through ventricles. However, ventricular enlargement during NPH may be caused by backflow into A previous study showed this reversal flow; present study, authors provide additional clinical data obtained patients with and supplement these computer simulations to better understand wall displacement emphasize its implications.Three 1 patient aqueductal stenosis underwent...
Patients with the familial form of cerebral cavernous malformations (CCMs) are haploinsufficient for CCM1, CCM2, or CCM3 gene. Loss corresponding CCM proteins increases RhoA kinase-mediated endothelial permeability in vitro, and mouse brains vivo. A prospective case-controlled observational study investigated whether human subjects show vascular hyperpermeability by dynamic contrast-enhanced quantitative perfusion magnetic resonance imaging, comparison cases without disease, lesional brain...
Hyperpermeability and iron deposition are 2 central pathophysiological phenomena in human cerebral cavernous malformation (CCM) disease. Here, we used novel MRI techniques to establish a relationship between these phenomena.Subjects with CCM disease (4 sporadic 17 familial) underwent imaging using the dynamic contrast-enhanced quantitative perfusion susceptibility mapping that measure hemodynamic factors of vessel leak deposition, respectively, previously demonstrated Regions interest...
Spectral computed tomography based on a photon-counting detector (PCD) attracts more and attentions since it has the capability to provide accurate identification quantitative analysis for biomedical materials. The limited number of photons within narrow energy bins leads imaging results low signal-noise ratio. existing supervised deep reconstruction networks CT are difficult address these challenges because is usually impossible acquire noise-free clinical images with clear structures as...
Semantic segmentation networks focus on the scene parsing of an unrestricted open scene. The typical architectures are stacks consisting convolutional layers, which used to extract semantic features. feature map dimension is sharply changed at sampling units for most networks, ensure effective propagation gradient in deep nets. In this article, we proposed a state-of-the-art network model named Fully Convolutional Pyramidal Networks (FC-PRNet), employs pyramidal residual structure change all...
Since December 2019, many patients in Wuhan sustained pneumonia with unknown causes.[1,2] A part of severe developed acute respiratory distress syndrome or septic shock, and even death. On January 7, 2020, Chinese researchers have for the first time detected a new coronavirus. 20, National Health Commission issued Announcement No. 1 which included coronavirus as Class B infectious disease, conducted epidemic prevention control according to disease. This kind has been named disease 2019...
Purpose Direct mapping of neuronal currents using MRI would have fundamental impacts on brain functional imaging. Previous reports indicated that the stimulus‐induced rotary saturation (SIRS) mechanism had best potential direct detection neural oscillations; however, it lacked high‐sensitivity level needed. In this study, a novel strategy is proposed in an effort to improve sensitivity. Methods our modified SIRS sequence, external oscillatory magnetic field used as excitation pulse place...
BACKGROUND:Spectral computed tomography (CT) has the capability to resolve energy levels of incident photons, which potential distinguish different material compositions. Although decomposition methods based on x-ray attenuation characteristics have good performance in dual- CT imaging, there are some limitations terms image contrast and noise levels. OBJECTIVE:This study focused multi-material spectral images a deep learning approach. METHODS:To classify quantify materials, we proposed...
Abstract Coronavirus disease 2019 (COVID-19) has brought huge losses to the world, and it remains a great threat public health. X-ray computed tomography (CT) plays central role in management of COVID-19. Traditional diagnosis with pulmonary CT images is time-consuming error-prone, which could not meet need for precise rapid COVID-19 screening. Nowadays, deep learning (DL) been successfully applied image analysis, assists radiologists workflow scheduling treatment planning patients methods...
Photon-counting micro computed tomography (micro-CT) offers new potential in preclinical imaging, particularly distinguishing materials. It becomes especially helpful when combined with contrast agents, enabling the differentiation of tumors from surrounding tissues. There are mainly two types agents market for micro-CT: small molecule-based and nanoparticle-based. However, despite their widespread use liver tumor studies, there is a notable gap research on application these commercially...
Laryngeal Tumor Grading is a challenge task for computer-aided clinical diagnosis (CACD), mainly because the nuclei in histopathological images have large differences shape and distribution, complex spatial relationship between nuclei. However, existing CNN-based tumor grading models cannot adaptively represent with variable morphology, lack of effective modeling semantic information nuclear location. Therefore, we propose an end-to-end network (DCA-DAFFNet) deformable convolution guided...