A5014360407- Sepsis Diagnosis and Treatment
- Radiomics and Machine Learning in Medical Imaging
- Machine Learning in Healthcare
- Acute Kidney Injury Research
- AI in cancer detection
- Cardiac Arrest and Resuscitation
- Lung Cancer Diagnosis and Treatment
- Music and Audio Processing
- Artificial Intelligence in Healthcare
- Pancreatic and Hepatic Oncology Research
- Blood properties and coagulation
- Speech and Audio Processing
- Traumatic Brain Injury and Neurovascular Disturbances
- Chaos-based Image/Signal Encryption
- Speech Recognition and Synthesis
- Trauma and Emergency Care Studies
- COVID-19 diagnosis using AI
- Cell Image Analysis Techniques
- Digital Imaging for Blood Diseases
- Hydrological Forecasting Using AI
- Neural Networks and Applications
- Hemostasis and retained surgical items
- Phonocardiography and Auscultation Techniques
- Brain Metastases and Treatment
- Hallucinations in medical conditions
Zhejiang University
2024-2025
Hong Kong Polytechnic University
2023-2024
Shenzhen Polytechnic
2024
Shenzhen University
2020-2023
Shenzhen University Health Science Center
2021-2023
Harbin Engineering University
2021
<p>Artificial Intelligence (AI) has transformed how we live and think, it will change practice medicine. With multimodal big data, can develop large medical models that enables what used to unimaginable, such as early cancer detection several years in advance effective control of virus outbreaks without imposing social burdens. The future is promising, are witnessing the advancement. That said, there challenges cannot be overlooked. For example, data generated often isolated difficult...
Monitoring of the coagulation function has applications in many clinical settings. Routine assays clinic are sample-consuming and slow turnaround. Microfluidics provides opportunity to develop that applicable point-of-care settings, but reported works required bulky sample pumping units or costly data acquisition instruments. In this work, we developed a microfluidic assay with simple setup easy operation. The device continuously generated droplets blood buffer mixture temporal development...
With the rapid development of speech synthesis and voice conversion technologies, Audio Deepfake has become a serious threat to Automatic Speaker Verification (ASV) system. Numerous countermeasures are proposed detect this type attack. In paper, we report our efforts combine self-supervised WavLM model Multi-Fusion Attentive classifier for audio deepfake detection. Our method exploits extract features that more conducive spoofing detection first time. Then, propose novel (MFA) based on...
Traditional scoring systems for patients' outcome prediction in intensive care units such as Oxygenation Saturation Index (OSI) and (OI) may not reliably predict the clinical prognosis of patients with acute respiratory distress syndrome (ARDS). Thus, none them have been widely accepted mortality ARDS. This study aimed to develop validate a method ARDS based on machine learning using Medical Information Mart Intensive Care (MIMIC-III) Telehealth Unit (eICU) Collaborative Research Database...
This study aimed to develop and validate a model for mortality risk stratification of intensive care unit (ICU) patients with acute kidney injury (AKI) using the machine learning technique.Eligible data were extracted from Medical Information Mart Intensive Care (MIMIC-III) database. Calibration, discrimination, classification prediction evaluated conventional scoring systems new algorithm. A 10-fold cross-validation was performed. The predictive models externally validated eICU database...
It is a critical challenge to diagnose leptomeningeal metastasis (LM), given its technical difficulty and the lack of typical symptoms. The existing gold standard diagnosing LM use positive cerebrospinal fluid (CSF) cytology, which consumes significantly more time classify cells under microscope.This study aims establish deep learning model cancer in CSF, thus facilitating doctors achieve an accurate fast diagnosis early stage.The laboratory Xijing Hospital provides 53,255 from 90 patients...
The mortality rate in the intensive care unit (ICU) is a key metric of hospital clinical quality. To enhance performance, many methods have been proposed for stratification patients' different risk categories, such as severity scoring systems and machine learning models. However, these make capturing time sequence information difficult, posing challenges to continuous assessment patient's during their stay. Therefore, we built predictive model that can predictions throughout stay obtain...
This study aimed to construct a mortality model for the risk stratification of intensive care unit (ICU) patients with sepsis by applying machine learning algorithm.Adult who were diagnosed during admission ICU extracted from MIMIC-III, MIMIC-IV, eICU, and Zigong databases. MIMIC-III was used development internal validation. The other three databases external Our proposed developed based on Extreme Gradient Boosting (XGBoost) algorithm. generalizability, discrimination, validation our...
In view of the fact that radiomics features have been reported as predictors immunotherapy to various cancers, this study aimed develop a prediction model determine response anti-programmed death-1 (anti-PD-1) therapy in esophageal squamous cell carcinoma (ESCC) patients from contrast-enhanced CT (CECT) features.Radiomic analysis images was performed retrospectively for image samples before and after anti-PD-1 treatment, efficacy results two different time node evaluations. A total 68 were...
Acute kidney injury (AKI) is associated with increased mortality in critically ill patients. Due to differences the etiology and pathophysiological mechanism, current AKI criteria put it an embarrassment evaluate clinical therapy prognosis. We aimed identify subphenotypes based on routinely collected data expose unique pathophysiologic patterns. A retrospective study was conducted Medical Information Mart for Intensive Care IV (MIMIC-IV) eICU Collaborative Research Database (eICU-CRD), a...
Acute kidney injury (AKI), a common condition on the intensive-care unit (ICU), is characterized by an abrupt decrease in function within few hours or days, leading to failure damage. Although AKI associated with poor outcomes, current guidelines overlook heterogeneity among patients this condition. Identification of subphenotypes could enable targeted interventions and deeper understanding injury's pathophysiology. While previous approaches based unsupervised representation learning have...
The exact definition of Acute kidney injury (AKI) for patients with traumatic brain (TBI) is unknown.To compare the power "Risk, Injury, Failure, Loss function, and End-stage disease" (RIFLE), Kidney Injury Network (AKIN), Creatinine kinetics (CK), Disease Improving Global Outcomes (KDIGO) to determine AKI incidence/stage their association in-hospital mortality rate TBI.This retrospective study collected data admitted intensive care unit neurotrauma from 2001 2012, 1648 were included....
Deep learning has been extensively used in various fields, such as phase imaging, 3D imaging reconstruction, unwrapping, and laser speckle reduction, particularly for complex problems that lack analytic models. Its data-driven nature allows implicit construction of mathematical relationships within the network through training with abundant data. However, a critical challenge practical applications is generalization issue, where trained on one dataset struggles to recognize an unknown target...
Computational histopathology is a fast emerging field which converts the traditional glass slide based department to new examination platform. Such paradigm shift also brings in silico computation field. Much research have been presented past decades on algorithm development for pathology image analysis. On other hand, comprehensive software platform with advanced visualization and capability, large developer community, flexible plugin mechanism, friendly transnational license, would be...
The novel coronavirus disease 2019 (COVID-19) pandemic is a global threat caused by the severe acute respiratory syndrome coronavirus-2.To develop and validate risk stratification tool for early prediction of intensive care unit (ICU) admission among COVID-19 patients at hospital admission.The training cohort included admitted to Wuhan Third Hospital. We selected 13 65 baseline laboratory results assess ICU risk, which were used model with random forest (RF) algorithm. A nomogram logistic...
Optical encryption is pivotal in information security, offering parallel processing, speed, and robust security. The simplicity compatibility of speckle‐based cryptosystems have garnered considerable attention. Yet, the predictable statistical distribution speckle optical fields’ characteristics can invite attacks, undermining these methods. proposed solution, a deep adversarial learning‐based modulation network (DeepSLM), disrupts strong intercorrelation grains. Utilizing unique encoding...
With fast developments in computational power and algorithms, deep learning has made breakthroughs been applied many fields. However, generalization remains to be a critical challenge, the limited capability severely constrains its practical applications. Hallucination issue is another unresolved conundrum haunting large models. By leveraging physical model of imaging through scattering media, we studied lack system response functions learning, identified cause, proposed universal solution....
With the rapid development of speech synthesis and voice conversion technologies, Audio Deepfake has become a serious threat to Automatic Speaker Verification (ASV) system. Numerous countermeasures are proposed detect this type attack. In paper, we report our efforts combine self-supervised WavLM model Multi-Fusion Attentive classifier for audio deepfake detection. Our method exploits extract features that more conducive spoofing detection first time. Then, propose novel (MFA) based on...
Abstract Pancreatoscopy plays a significant role in the diagnosis and treatment of pancreatic diseases. However, risk pancreatoscopy is remarkably greater than that other endoscopic procedures, such as gastroscopy bronchoscopy, owing to its severe invasiveness. In comparison, virtual (VP) has shown notable advantages. because low resolution current computed tomography (CT) technology small diameter duct, VP limited clinical use. this study, an optimal path algorithm super-resolution...
Septic patients admitted to the intensive care unit (ICU) are highly susceptible acute kidney injury (AKI), which leads reduced survival in these patients. It is thus necessary develop a model that can predict risk of AKI septic real time. Although continuous or near-continuous assessment likely necessary, few models have been designed for this purpose. Therefore, we constructed continuously sepsis-induced ICU. Our proposed optimally achieved an area under receiver operating characteristic...
Arithmetic Optimization Algorithm (AOA) is a new-proposed meta-heuristic algorithm motivated by the distribution behavior of leading arithmetic operators. It proved to perform well in optimization processes complex engineering design problems. AOA has slow convergence rate and tends fall into local minima multimodal This paper proposes an based on particle energy diffusion. On one hand, introduction increases chance updating particle's position, which not gained satisfying solution,...