A5101931857- Medical Imaging Techniques and Applications
- Advanced X-ray and CT Imaging
- Advanced X-ray Imaging Techniques
- Microfluidic and Capillary Electrophoresis Applications
- Machine Fault Diagnosis Techniques
- Advanced MRI Techniques and Applications
- Radiation Dose and Imaging
- Innovative Microfluidic and Catalytic Techniques Innovation
- Anomaly Detection Techniques and Applications
- Radiomics and Machine Learning in Medical Imaging
- Advanced Electron Microscopy Techniques and Applications
- Hydrocarbon exploration and reservoir analysis
- IoT and Edge/Fog Computing
- Advanced Radiotherapy Techniques
- Satellite Communication Systems
- Fault Detection and Control Systems
- Age of Information Optimization
- Paleontology and Stratigraphy of Fossils
- Non-Destructive Testing Techniques
- Digital Radiography and Breast Imaging
- Geochemistry and Elemental Analysis
- Analytical Chemistry and Sensors
- Neural Networks and Reservoir Computing
- Space Satellite Systems and Control
- Green IT and Sustainability
Hubei University of Chinese Medicine
2025
Shenzhen Institutes of Advanced Technology
2011-2024
Chinese Academy of Sciences
2012-2024
Nanjing University of Chinese Medicine
2024
Shanghai Medical College of Fudan University
2024
Beijing University of Posts and Telecommunications
2021-2023
China University of Geosciences
2023
Anhui University
2022
Xi'an Jiaotong University
2020-2021
University of Chinese Academy of Sciences
2019-2020
Recent progress on intelligent fault diagnosis (IFD) has greatly depended deep representation learning and plenty of labeled data. However, machines often operate with various working conditions or the target task different distributions collected data used for training (the domain shift problem). Besides, newly test in are usually unlabeled, leading to unsupervised transfer based (UDTL-based) IFD problem. Although it achieved huge development, a standard open source code framework as well...
Existing data-driven fault diagnosis methods assume that the label sets of training data and test are consistent, which is usually not applicable for real applications since modes occur in phase unpredictable. To address this problem, open set (OSFD), where consists a portion some unknown classes, studied article. Considering changeable operating conditions machinery, OSFD tasks further divided into shared-domain (SOSFD) cross-domain (COSFD) For SOSFD, 1-D convolutional neural networks...
In recent years, health risks concerning high-dose x-ray radiation have become a major concern in dental computed tomography (CT) examinations. Therefore, adopting low-dose (LDCT) technology has focus the CT imaging field. One of these LDCT technologies is downsampling data acquisition during processes. However, reducing dose can adversely affect image quality by introducing noise and artifacts resultant that compromise diagnostic information. this paper, we propose an artifact correction...
Abstract Objectives This study aims to decrease the scan time and enhance image quality in pediatric total-body PET imaging by utilizing multimodal artificial intelligence techniques. Methods A total of 270 patients who underwent PET/CT scans with a uEXPLORER at Sun Yat-sen University Cancer Center were retrospectively enrolled. 18 F-fluorodeoxyglucose ( F-FDG) was administered dose 3.7 MBq/kg an acquisition 600 s. Short-term images (acquired within 6, 15, 30, 60 150 s) obtained truncating...
Traditional deep learning assumes that the probability distributions of test set and training are exactly same. When samples from different working conditions, distribution gap between conditions might dramatically lower accuracy. To address this problem, domain adaptation (DA) via joint using sets is adopted to increase final The disadvantage DA it requires participate in phase. However, sometimes impossible collect beforehand. solve article proposes a conditional adversarial generalization...
Recently, the paradigm of computed tomography (CT) reconstruction has shifted as deep learning technique evolves. In this study, we proposed a new convolutional neural network (called ADAPTIVE-NET) to perform CT image directly from sinogram by integrating analytical domain transformation knowledge.In ADAPTIVE-NET, specific layer with constant weights was customized transform onto via back-projection. With framework, feature extractions were performed simultaneously on both and domain. The...
Abstract The rapid development of generative artificial intelligence (GenAI) has generated significant economic and social value, alongside risks to user privacy. For this purpose, study investigates privacy protection in human-AI interaction by employing a combined approach evolutionary game system dynamics. A three-party model was developed analyze the interactive effects evolution strategies among government, GenAI company, users. Sensitivity analysis through dynamics simulations...
Background: Reducing the radiation tracer dose and scanning time during positron emission tomography (PET) imaging can reduce cost of tracer, motion artifacts, increase efficiency scanner. However, reconstructed images to be noisy. It is very important reconstruct high-quality with low-count (LC) data. Therefore, we propose a deep learning method called LCPR-Net, which used for directly reconstructing full-count (FC) PET from corresponding LC sinogram Methods: Based on framework generative...
The aim of this study is to demonstrate the feasibility removing image Moire artifacts caused by system inaccuracies in grating-based x-ray interferometry imaging via convolutional neural network (CNN) technique. Instead minimizing these inconsistencies between acquired phase stepping data certain optimized signal retrieval algorithms, our newly proposed CNN-based method reduces image-domain a learned post-processing procedure. To ease training preparations, we propose synthesize them with...
Together, the Yinggehai and Qiongdongnan basins have received a large amount of terrigenous sediments, but provenance evolution Cenozoic sediments in two remains disputable. Combined with previous studies basins, elemental geochemistry Oligocene to Pliocene sediment samples junction area were analyzed explore tectonic implications, parent rock characteristics, during Cenozoic. The results reveal that all derived from continental island arc passive margin settings. light REE enrichment stable...
Studying the accumulation rules of organic matter (OM) in paleo-ocean sediments can not only enhance our understanding how OM becomes enriched ancient oceans but also provide guidance for exploration shale gas unconventional strata. A breakthrough has been made early Cambrian Qiongzhusi Formation South China. However, less attention paid to intraplatform basin Yangtze Platform, and factors controlling enrichment this special region remain unclear. This study focuses on a continuous drilling...
Background: Lymph node retrieval deficiency can lead to understagement and postoperative cancer recurrence, it is crucial establish the standard number of retrieved lymph nodes (rLNs) negative (nLNs) for patients undergoing gastrectomy. Methods: Patients who has gastric adenocarcinoma underwent either radical subtotal gastrectomy (RSG) or total (RTG) between 2000 2022 were retrospectively included. The authors utilized restricted cubic spline (RCS) analysis determine ideal threshold rLNs...
X-ray radiation is harmful to human health. Thus, obtaining a better reconstructed image with few projection view constraints major challenge in the computed tomography (CT) field reduce dose. In this study, we proposed and tested new algorithm that combines penalized weighted least-squares using total generalized variation (PWLS-TGV) dictionary learning (DL), named PWLS-TGV-DL address challenge. We first presented evaluated it through both data simulation physical experiments. then analyzed...
For conventional computed tomography (CT) image reconstruction tasks, the most popular method is so-called filtered-back-projection (FBP) algorithm. In it, acquired Radon projections are usually filtered first by a ramp kernel before back-projected to generate CT images. this work, as contrary, we realized idea of image-domain backproject-filter (BPF) using deep learning techniques for time. With properly designed convolutional neural network (CNN), preliminary results demonstrate that it...
Total-body dynamic positron emission tomography (dPET) imaging using 18 F-fluorodeoxyglucose (18 F-FDG) has received widespread attention in clinical oncology. However, the conventionally required scan duration of approximately 1 h seriously limits application and promotion this technique. In study, we investigated possibility feasibility shortening total-body to 30 min post-injection (PI) with help a novel Patlak data processing algorithm for accurate Ki estimations tumor lesions.Total-body...
Radiation risk from computed tomography (CT) is always an issue for patients, especially those in clinical conditions which repeated CT scanning required. For patients undergoing scanning, a low-dose protocol, such as sparse often used, and consequently, advanced reconstruction algorithm also needed.To develop novel used sparse-view associated with the prior image.A method based on information of normal-dose image (PI-NDI) involving transformed model attenuation coefficients object to be...
LIF detection often requires labeling of analytes with fluorophores; and fast fluorescent derivatization is valuable for high-throughput analysis flow-gated CE. Here, we report a fluorescein-labeling scheme amino acid neurotransmitters, which were then rapidly separated detected in This was based on the reaction between primary amines o-phthalaldehyde presence thiol, 2-((5-fluoresceinyl)aminocarbonyl)ethyl mercaptan (FACE-SH). The short time (<30 s) suited on-line mixing that directly...