A5051234199- Biomedical Text Mining and Ontologies
- Topic Modeling
- Electronic Health Records Systems
- Natural Language Processing Techniques
- Semantic Web and Ontologies
- Information Retrieval and Search Behavior
- Advanced Vision and Imaging
- Advanced Text Analysis Techniques
- Radiomics and Machine Learning in Medical Imaging
- Machine Learning in Healthcare
- Artificial Intelligence in Healthcare and Education
- Advanced Image Processing Techniques
- AI in cancer detection
- Innovation in Digital Healthcare Systems
- Advanced Optical Sensing Technologies
- Artificial Intelligence in Healthcare
- Technology and Data Analysis
- Service-Oriented Architecture and Web Services
- Emergency and Acute Care Studies
- Traffic Prediction and Management Techniques
- Domain Adaptation and Few-Shot Learning
- Internet of Things and Social Network Interactions
- Pneumonia and Respiratory Infections
- Genomics and Chromatin Dynamics
- Advanced Image and Video Retrieval Techniques
Seoul National University
2016-2025
New Generation University College
2008-2025
Gwangju Institute of Science and Technology
2023
Kyungpook National University
2023
Hannam University
2022
Advisory Board Company (United States)
2021
Weatherford College
2021
Soonchunhyang University
2021
Check Point (Israel)
2021
Bio-Medical Science (South Korea)
2020
: Dental panoramic radiographs (DPRs) provide information required to potentially evaluate bone density changes through a textural and morphological feature analysis on mandible. This study aims the discriminating performance of deep convolutional neural networks (CNNs), employed with various transfer learning strategies, classification specific features osteoporosis in DPRs. For objective labeling, we collected dataset containing 680 images from different patients who underwent both...
Although renal hyperfiltration (RHF) or an abnormal increase in GFR has been associated with many lifestyles and clinical conditions, including diabetes, its consequence is not clear. RHF frequently considered to be the result of overestimating true subjects muscle wasting. To evaluate association between mortality, 43,503 adult Koreans who underwent voluntary health screening at Seoul National University Hospital March 1995 May 2006 baseline GFR≥60 ml/min per 1.73 m2 were followed up for...
Pathology Artificial Intelligence Platform (PAIP) is a free research platform in support of pathological artificial intelligence (AI). The main goal the to construct high-quality pathology learning data set that will allow greater accessibility. PAIP Liver Cancer Segmentation Challenge, organized conjunction with Medical Image Computing and Computer Assisted Intervention Society (MICCAI 2019), first image analysis challenge apply datasets. was evaluate new existing algorithms for automated...
Abstract Despite substantial advances in disease genetics, studies to date have largely focused on individuals of European descent. This limits further discoveries novel functional genetic variants other ethnic groups. To alleviate the paucity East Asian population genome resources, we established Korean Variant Archive 2 (KOVA 2), which is composed 1896 whole-genome sequences and 3409 whole-exome from healthy ethnicity. largest database date, surpassing 1909 deposited gnomAD. The KOVA...
One major challenge associated with lung cancer organoids (LCOs) is their predominant derivation from surgical specimens of patients early-stage cancer. However, advanced cancer, who are in need chemotherapy, often cannot undergo surgery. Therefore, there an urgent to successfully generate LCOs biopsy specimens. Conventional techniques, such as transthoracic needle and forceps biopsy, only yield small amounts tissue, resulting a low success rate for culturing samples. Furthermore, potential...
The steadily increasing number of medical images places a tremendous burden on doctors, who toned to read and write reports. If an image captioning model could generate drafts the reports from corresponding images, workload doctors would be reduced, thereby saving time expenses. aim this study was develop chest x-ray that considers differences between patient normal uses hierarchical long short-term memory (LSTM) or transformer as decoder We investigated which feature representation method...
Patient clinical data are distributed and often fragmented in heterogeneous systems, therefore the need for information integration is a key to reliable patient care. Once orderly integrated readily available, problems accessing data, well-known difficulties of adopting mobile health system, resolved. This paper proposes system (MobileMed), which integrates across sources makes them accessible through devices. The consists four main components: smart interface, an HL7 message server (HMS),...
We introduce a novel method to automatically adjust camera exposure for image processing and computer vision applications on mobile robot platforms. Because most algorithms rely heavily low-level features that are based mainly local gradient information, we consider quantity can determine the proper level, allowing capture important in manner robust illumination conditions. then extend this concept multi-camera system present new control algorithm achieve both brightness consistency between...
The emergency department (ED) triage system to classify and prioritize patients from high risk less urgent continues be a challenge.This study, comprising 80,433 patients, aims develop machine learning algorithm prediction model of critical care outcomes for adult using information collected during ED compare the performance with that baseline Korean Triage Acuity Scale (KTAS).To predict need care, we used 13 predictors information: age, gender, mode arrival, time interval between onset...
Microsatellite instability (MSI) refers to alterations in the length of simple repetitive genomic sequences. MSI status serves as a prognostic and predictive factor colorectal cancer. The MSI-high is good stage II/III cancer, predicts lack benefit adjuvant fluorouracil chemotherapy II cancer but response immunotherapy IV Therefore, determining patients with important for identifying appropriate treatment protocol. In Pathology Artificial Intelligence Platform (PAIP) 2020 challenge,...
Abstract Single-cell transcriptomics enables the study of cellular heterogeneity, but current unsupervised strategies make it challenging to associate individual cells with sample conditions. We propose scMILD, a weakly supervised learning framework based on Multiple Instance Learning, which leverages sample-level labels identify condition-associated cell subpopulations. scMILD employs dual-branch architecture perform classification and cell-level representation simultaneously. validated...
Developing large language models (LLMs) in biomedicine requires access to high-quality training and alignment tuning datasets. However, publicly available Korean medical preference datasets are scarce, hindering the advancement of LLMs. This study constructs evaluates efficacy Medical Preference Dataset (KoMeP), an dataset constructed with automated pipeline, minimizing high costs human annotation. KoMeP was generated using DAHL score, hallucination evaluation metric. Five LLMs (Dolly-v2-3B,...
<title>Abstract</title> Transposable elements (TEs) are essential genomic entities that play the roles of eukaryotic genome regulators and involved in controlling gene expression patterns, cell-type specialization, diseases. Recent improvements single-cell assay for transposase-accessible chromatin sequencing (scATAC-seq) have enhanced accessibility investigation with cell type specificity. However, existing techniques do not address locus-specific activity TEs. Here, we present scTELL...
<title>Abstract</title> Efficient extraction of structured information from unstructured radiology reports remains a critical challenge in healthcare. We introduce the Radiology Report Information Extraction Framework (RRIEF), privacy-preserving approach utilizing parameter-efficient fine-tuning open-source large language models (LLMs). validated RRIEF across chest X-ray (CXR), mammography, and coronary CT angiography (CCTA) reports, evaluating its performance against specialized methods...
Pathology reports are written in free-text form, which precludes efficient data gathering. We aimed to overcome this limitation and design an automated system for extracting biomarker profiles from accumulated pathology reports. designed a new model representing knowledge. The parses immunohistochemistry based on "slide paragraph" unit defined as set of findings obtained the same tissue slide. parsed using context-free grammar immunohistochemistry, tree-like structure surgical pathology....
The number of international benchmarking competitions is steadily increasing in various fields machine learning (ML) research and practice. So far, however, little known about the common practice as well bottlenecks faced by community tackling questions posed. To shed light on status quo algorithm development specific field biomedical imaging analysis, we designed an survey that was issued to all participants challenges conducted conjunction with IEEE ISBI 2021 MICCAI conferences (80 total)....
A seasonal variation of glucose homeostasis in humans has been reported various geographic regions.In this study, we examined variations hemoglobin A1c (HbA1c) patients with type 2 diabetes living Korea.We analyzed 57,970 HbA1c values from 4,191 and the association these ambient temperature for 3.5 yr.Overall, exhibited its highest February to March lowest September October (coefficient cos t = -0.0743,P 0.058) difference between peak nadir a year was 0.16%-0.25%.A statistically significant...