A5067147327- Photoacoustic and Ultrasonic Imaging
- Thermography and Photoacoustic Techniques
- Advanced X-ray and CT Imaging
- Anomaly Detection Techniques and Applications
- Mobile Agent-Based Network Management
- Atomic and Subatomic Physics Research
- Artificial Intelligence in Healthcare
- Optical Imaging and Spectroscopy Techniques
- IoT and Edge/Fog Computing
- Advanced MRI Techniques and Applications
- Topic Modeling
- Multi-Agent Systems and Negotiation
- Video Analysis and Summarization
- Neural Networks and Applications
- Recommender Systems and Techniques
- Ultrasound and Hyperthermia Applications
- Multimedia Communication and Technology
- Atmospheric and Environmental Gas Dynamics
- Service-Oriented Architecture and Web Services
- Evolutionary Algorithms and Applications
- Face and Expression Recognition
- Radiomics and Machine Learning in Medical Imaging
- Metaheuristic Optimization Algorithms Research
- Cardiac Imaging and Diagnostics
- Wireless Sensor Networks and IoT
Mitre (United States)
2019-2024
Xi’an Jiaotong-Liverpool University
2011-2023
George Mason University
2019-2023
University of Virginia
2014-2022
Northeast Electric Power University
2022
Guangdong University of Technology
2021
Conference Board
2021
Brunel University of London
2007
National University of Singapore
2002-2006
Photoacoustic imaging is an emerging modality that based upon the photoacoustic effect. In tomography (PAT), induced acoustic pressure waves are measured by array of detectors and used to reconstruct image initial distribution. A common challenge faced in PAT can only be sparsely sampled. Reconstructing sampled data using standard methods results severe artifacts obscure information within image. We propose a modified convolutional neural network (CNN) architecture termed Fully Dense UNet...
Photoacoustic tomography (PAT) is a nonionizing imaging modality capable of acquiring high contrast and resolution images optical absorption at depths greater than traditional techniques. Practical considerations with instrumentation geometry limit the number available acoustic sensors their view target, which result in significant image reconstruction artifacts degrading quality. Iterative methods can be used to reduce but are computationally expensive. In this work, we propose novel deep...
Conventional reconstruction methods for photoacoustic images are not suitable the scenario of sparse sensing and geometrical limitation. To overcome these challenges enhance quality reconstruction, several learning-based have recently been introduced tomography reconstruction. The goal this study is to compare systematically evaluate proposed modified networks image Specifically, post-processing model-based learned iterative investigated. In addition comparing differences inherently brought...
Simulation tools for photoacoustic wave propagation have played a key role in advancing imaging by providing quantitative and qualitative insights into parameters affecting image quality. Classical methods numerically solving the equation relies on fine discretization of space can become computationally expensive large computational grids. In this work, we apply Fourier Neural Operator (FNO) networks as fast data-driven deep learning method 2D homogeneous medium. Comparisons between FNO...
Idiopathic pulmonary fibrosis, a pattern of interstitial lung disease, is often clinically unpredictable in its progression. This paper presents hyperpolarized Xenon-129 chemical shift imaging as noninvasive, nonradioactive method probing physiology well anatomy to monitor subtle changes subjects with IPF. Twenty subjects, nine healthy and eleven IPF, underwent HP Xe-129 ventilation MRI 3D-SBCSI. Spirometry was performed on all before imaging, DLCO hematocrit were measured IPF after imaging....
Iterative reconstruction has demonstrated superior performance in medical imaging under compressed, sparse, and limited-view sensing scenarios. However, iterative algorithms are slow to converge rely heavily on hand-crafted parameters achieve good performance. Many iterations usually required reconstruct a high-quality image, which is computationally expensive due repeated evaluations of the physical model. While learned approaches such as model-based learning (MBLr) can reduce number...
3D Single-breath Chemical Shift Imaging (3D-SBCSI) is a hybrid MR-spectroscopic imaging modality that uses hyperpolarized xenon-129 gas (Xe-129) to differentiate lung diseases by probing functional characteristics. This study tests the efficacy of 3D-SBCSI in differentiating physiology among pulmonary diseases. A total 45 subjects—16 healthy, 11 idiopathic fibrosis (IPF), 13 cystic (CF), and 5 chronic obstructive disease (COPD)—were given 1/3 forced vital capacity (FVC) Xe-129, inhaled for...
In photoacoustic tomography (PAT), the acoustic pressure waves produced by optical excitation are measured an array of detectors and used to reconstruct image. Sparse spatial sampling limited-view detection two common challenges faced in PAT. Reconstructing from incomplete data using standard methods results severe streaking artifacts blurring. We propose a modified convolutional neural network (CNN) architecture termed Dense Dilation UNet (DD-UNet) for correcting 3D The DD-Net leverages...
Objective We quantified inflammatory burden in rheumatoid arthritis (RA) synovial tissue by using computer vision to automate the process of counting individual nuclei hematoxylin and eosin images. Methods adapted applied algorithms quantify density (count per unit area tissue) on from arthroplasty samples. A pathologist validated algorithm results labeling images that were mislabeled or missed algorithm. Nuclei was compared with other measures RA inflammation such as semiquantitative...
Simulation tools for photoacoustic wave propagation have played a key role in advancing imaging by providing quantitative and qualitative insights into parameters affecting image quality. Classical methods numerically solving the equation relies on fine discretization of space can become computationally expensive large computational grids. In this work, we apply Fourier Neural Operator (FNO) networks as fast data-driven deep learning method 2D homogeneous medium. Comparisons between FNO...
Although bearing fault diagnosis methods based on deep learning are very popular in recent years and a lot of brilliant results have been achieved, they assume that the distribution training samples is same with test samples. However, working condition variable, labeling tags for all data time-consuming laborious. In order to solve problem lacking labeled cross domain scenario, novel adaptation transfer method adversarial network proposed. this method, convolutional neural (CNN) used extract...
To avoid confusion, all abbreviations "GQL" throughout the text should be standardized as "GraphQL",
A software component is defined as a unit of composition with contractually specified interfaces and explicit dependencies that may be independently deployed. Components form generic, re-usable building blocks, which can composed into applications deployed by third parties. good model therefore must seek to minimize implicit in order maximize re-use composability. The benefits models have led their widespread application the area networked embedded systems particularly Wireless Sensor...
Practitioners and researchers often found the intrinsic representations of high-dimensional problems has much fewer independent variables. However such structure may not be easily discovered due to noises other factors. A supervised transformation scheme RST is proposed transform features into lower dimensional spaces for classification tasks. The algorithm recursively selectively transforms guided by output We compared performance linear classifier random forest on original data sets, sets...