A5100693794- Dental Radiography and Imaging
- Antimicrobial Resistance in Staphylococcus
- Sepsis Diagnosis and Treatment
- Machine Learning in Healthcare
- Medical Imaging and Analysis
- Facial Trauma and Fracture Management
- Advanced Image Processing Techniques
- AI in cancer detection
- Image and Video Quality Assessment
- Digital Imaging in Medicine
- Biomedical Text Mining and Ontologies
- Gastrointestinal Bleeding Diagnosis and Treatment
- COVID-19 diagnosis using AI
- Radiomics and Machine Learning in Medical Imaging
- Image and Signal Denoising Methods
- Lung Cancer Diagnosis and Treatment
- Ultrasound and Hyperthermia Applications
- Chemical synthesis and pharmacological studies
- Medical Imaging Techniques and Applications
- Mesenchymal stem cell research
- Eicosanoids and Hypertension Pharmacology
- Forensic Toxicology and Drug Analysis
- Functional Brain Connectivity Studies
- Hydrological Forecasting Using AI
- Facial Nerve Paralysis Treatment and Research
Kangwon National University
2005-2025
Korea Institute of Industrial Technology
2024
National Institutes of Health Clinical Center
2023
Dongguk University Ilsan Hospital
2023
University of California, Irvine
2012
Korea Institute of Science and Technology
2009
Korea University
2003
Worldwide, sepsis is the leading cause of death in hospitals. If mortality rates patients with can be predicted early, medical resources allocated efficiently. We constructed machine learning (ML) models to predict a hospital emergency department.
Facial bone fractures must be diagnosed and treated as early possible to avoid complications sequelae. CT images need analyzed detect fractures, but the analysis is time-consuming, enough specialists are not available analyze them. Many classification object detection studies being conducted address these issues. The ability of classification-based pinpoint exact location limited. Object detection-based research, by contrast, problematic because shape a fracture ambiguous. We propose...
The human facial bone is made up of many complex structures, which makes it challenging to accurately analyze fractures. To address this, we developed advanced image analysis software segments and quantifies spaces between fractured bones in CT images at the pixel level. This study used 3D scans from 1766 patients who had fractures a university hospital 2014 2020. Our solution included segmentation model focuses on identifying gaps created by However, training this required costly...
Wireless capsule endoscopy (WCE) has significantly advanced the diagnosis of gastrointestinal (GI) diseases by allowing for non-invasive visualization entire small intestine. However, machine learning-based methods organ classification in WCE often rely on color information, leading to decreased performance when obstacles such as food debris are present. This study proposes a novel model that integrates convolutional neural networks (CNNs) and long short-term memory (LSTM) analyze multiple...
In this paper, we present a novel postprocessing technique based on the theory of projection onto convex sets (POCS) in order to reduce blocking artifacts digital high definition television (HDTV) images. By detecting and eliminating undesired high-frequency components, mainly caused by artifacts, propose new smoothness constraint set (SCS) its operator DCT domain. addition, an improved quantization (QCS) using correlation coefficients between adjacent blocks. proposed technique, range QCS...
The incidence of facial fractures is on the rise globally, yet limited studies are addressing diverse forms present in 3D images. In particular, due to nature fracture, direction which bone vary, and there no clear outline, it difficult determine exact location fracture 2D Thus, image analysis required find area, but needs heavy computational complexity expensive pixel-wise labeling for supervised learning. this study, we tackle problem reducing burden increasing accuracy localization by...
Previously, we reported that 4-amino-2H-benzo[h]chromen-2-one (ABO) and 4-amino-7,8,9,10-tetrahydro-2H-benzo[h]chromen-2-one (ATBO) analogues, which were developed from the lead natural product neo-tanshinlactone, are potent cytotoxic agents. In order to improve on their water solubility, diamino analogues related salts designed. All synthesized compounds assayed for cytotoxicity, selected evaluated in vivo anti-mammary epithelial proliferation activity wild-type mice predisposed mammary...
Early detection of lung nodules is essential for preventing cancer. However, the number radiologists who can diagnose limited, and considerable effort time are required. To address this problem, researchers investigating automation deep-learning-based nodule detection. deep learning requires large amounts data, which be difficult to collect. Therefore, data collection should optimized facilitate experiments at beginning studies. We collected chest computed tomography scans from 515 patients...
Vancomycin pharmacokinetics are highly variable in patients with critical illnesses, and clinicians commonly use population pharmacokinetic (PPK) models based on a Bayesian approach to dose. However, these population-dependent, may only sometimes meet the needs of individual patients, used by experienced as reference for making treatment decisions. To assist real-world clinicians, we developed deep learning-based decision-making system that predicts vancomycin therapeutic drug monitoring...
Hematopoietic stem cell transplantation using bone marrow as the graft source is a common treatment for hematopoietic malignancies and disorders. For allogeneic transplants, processing of requires depletion ABO-mismatched red blood cells (RBCs) to avoid transfusion reactions. Here authors tested use an automated closed system depleting RBCs from compared results semi-automated platform that more commonly used in transplant centers today. The found fully Sepax instrument (Cytiva, Marlborough,...
Osteoporosis is a skeletal disease that difficult to identify in advance of symptoms. Existing screening methods, such as dual-energy X-ray absorptiometry, are only used for specific purpose due cost and safety reasons once symptoms develop. Early detection osteopenia osteoporosis using other modalities relatively frequent examinations helpful terms early treatment cost. Recently, many studies have proposed deep learning-based diagnosis methods various achieved outstanding results. However,...
Pressure ulcers (PUs) are a prevalent skin disease affecting patients with impaired mobility and in high-risk groups. These increase patients’ suffering, medical expenses, burden on staff. This study introduces clinical decision support system verifies it for predicting real-time PU occurrences within the intensive care unit (ICU) by using MIMIC-IV in-house ICU data. We develop various machine learning (ML) deep (DL) models real time validate Kangwon National University Hospital (KNUH)...
In this paper, we propose a new post-processing method, based on the theory of projection onto convex sets (POCS) to reduce in digital high definition television (HDTV) decoded images. We smoothness constraint set (SCS) and its operator wavelet transform (WT) domain remove unnecessary high-frequency components caused by blocking artifacts. also method find preserve original frequency image edge. Experimental results show that proposed can not only achieve significantly enhanced subjective...
This operative report documents the details of a surgery. Standardization medical terminology for written in free text is significant performing research and establishing insurance systems by accurately sharing information on treatment. However, standardization reports labor-intensive task that has risk induced errors. We have proposed concatenation bidirectional encoder representations from transformers (ConBERT) model predicting International Classification Disease-9 code using diagnosis...
Abstract Herein, we developed an innovative transdermal drug delivery system utilizing electrospinning technology. The employs reactive oxygen species (ROS)-responsive drug-loaded ferrocene-based polymer fibers (FPF), with ascorbyl tetraisopalmitate (AT) as the loaded drug. These offer superior protection from external factors, exhibit a high loading efficiency (~ 70 wt%), and enable controlled release under ROS conditions, such wounds. transition hydrophobic to hydrophilic state, triggered...
This study introduces a deep-learning-based automatic sleep scoring system to detect apnea using single-lead electrocardiography (ECG) signal, focusing on accurately estimating the apnea-hypopnea index (AHI). Unlike other research, this work emphasizes AHI estimation, crucial for diagnosis and severity evaluation of apnea. The suggested model, trained 1465 ECG recordings, combines deep-shallow fusion network detection (DSF-SANet) gated recurrent units (GRUs) analyze signals at 1-min...
<title>Abstract</title> The human facial bone is made up of many complex structures, which makes it challenging to accurately analyze fractures. To address this, we developed an advanced image analysis software that segments and quantifies spaces between fractured bones in CT images at the pixel level. This study used 3D scans from 1,766 patients who had fractures a university hospital 2014 2020. Our solution included segmentation model focuses on identifying gaps created by However,...
<sec> <title>BACKGROUND</title> An operative report is a document containing the details of surgery. The standardization medical terminology appearing in that written free text important for research and establishment insurance systems by allowing accurate sharing information on treatment. However, reports labor-intensive task with risk induced errors. </sec> <title>OBJECTIVE</title> We propose concatenated bidirectional encoder representation from transformers (ConBERT) model predicting...
Abstract Background Deep learning-based technologies are widely used in gastroenterology, especially for the lesion detections during endoscopic examination. However, there have been limited studies on localization of wireless capsule endoscopy (WCE). We propose an algorithm classifying organs and predicting gastric small bowel transit time WCE Methods collected data from 126 patients, using PillCam™ SB3 (Medtronic, Minneapolis, MN, USA) at two university-affiliated hospitals. The dataset...
<title>Abstract</title> This study focuses on predicting the prognosis of acute ischemic stroke patients with focal neurologic symptoms using a combination diffusion-weighted magnetic resonance imaging (DWI) and clinical information. The primary outcome is poor functional defined by modified Rankin Scale (mRS) score 3-6 after 3 months stroke. Employing nnUnet for DWI lesion segmentation, utilizes both multi-task multi-modality methodologies, integrating data prediction. Integrating two...