- Educational Systems and Policies
- Structural Behavior of Reinforced Concrete
- Advanced Neural Network Applications
- Advanced Image and Video Retrieval Techniques
- Education and Learning Interventions
- Structural Load-Bearing Analysis
- Plasma Diagnostics and Applications
- Plasma Applications and Diagnostics
- Simulation Techniques and Applications
- Vibrio bacteria research studies
- Electrohydrodynamics and Fluid Dynamics
- Metal and Thin Film Mechanics
- Image Retrieval and Classification Techniques
- Seismic Performance and Analysis
- Bacterial biofilms and quorum sensing
- Distributed and Parallel Computing Systems
- Education, Safety, and Science Studies
- Housing Market and Economics
- Network Security and Intrusion Detection
- Machine Learning and Data Classification
- Neural Networks and Applications
- Fire effects on concrete materials
- Aquaculture disease management and microbiota
- Graph Theory and Algorithms
- Head and Neck Cancer Studies
Keimyung University
2020-2025
Seoul National University
2015-2024
Korea Advanced Institute of Science and Technology
2024
Chung-Ang University
2020-2024
Pohang Iron and Steel (South Korea)
2020-2023
UNSW Sydney
2023
Inha University
2023
Cheongju National University of Education
2023
Dankook University
2022
University of Michigan
2022
Intradialytic hypotension has high clinical significance. However, predicting it using conventional statistical models may be difficult because several factors have interactive and complex effects on the risk. Herein, we applied a deep learning model (recurrent neural network) to predict risk of intradialytic timestamp-bearing dataset.We obtained 261,647 hemodialysis sessions with 1,600,531 independent timestamps (i.e., time-varying vital signs) randomly divided them into training (70%),...
Few-shot object detection aims to detect instances of specific categories in a query image with only handful support samples. Although this takes less effort than obtaining enough annotated images for supervised detection, it results far inferior performance compared the conventional methods. In paper, we propose meta-learning-based approach that con-siders unique characteristics each sample. Rather simply averaging information samples generate single prototype per category, our method can...
An experimental study was performed to investigate the structural behavior of thin-walled rectangular concrete-filled tubular (RCFT) columns. This mainly focused on effects a high-strength steel slender section overall eccentric compression capacity. The test parameters included width-to-thickness ratio plates, yield strength and use stiffeners. Five specimens were tested under axial loading. In slender-section specimens, despite early local buckling, significant postbuckling reserve...
Abstract Geographic location and neighbourhood attributes are major contributors to residential property values. Automated valuation models (AVM) often use hedonic pricing with in the form of numeric categorical variables. This paper proposed a novel approach automated using machine learning model convolutional neural network (CNN), fully connected layers In this study we compare results fused model, which treat geographical data as an input performance baseline only numerically or...
Abstract Typhoon Soulik decayed rapidly via two‐way interaction with the northern East China Sea, extratropical shelf region, before landing on Korean Peninsula 23 August 2018. In where water column is strongly stratified warm surface and cold subsurface waters, a large wake emerged intense sea cooling, which primarily caused by vertical mixing in column. This abrupt cooling increased downward enthalpy flux, mostly form of latent heat flux. The slow translation speed strong thermal...
Previous research has investigated non-pharmacological digital therapeutic interventions to improve compliance and reduce side effects in attention-deficit/hyperactivity disorder (ADHD) medication treatments for children. This study focuses on validating the of game-based intervention content enhancing working memory concentration. It tracks quantitative changes evaluate improvements concentration when is used as adjunct therapy alongside children with ADHD. Thirty participated; one group...
The profiles of the magnetic field component RF field, impedance characteristics and plasma parameters (electron density, electron temperature potential) a 13.56 MHz, planar inductively coupled argon were studied in presence an external compared with those non-magnetized discharge. An efficient power transfer stable matching low Q factor system could be obtained application relatively (5 - 20 G) pressure range 0.5 10 mTorr. measurement revealed that wave excitation was responsible for such...
ABSTRACT Introduction Exercise‐induced apelin as a myokine is believed to play role in the improvement of type 2 diabetes mellitus (T2DM) and capillarization. In this study, we evaluated association between exercise‐induced muscle Methods Zucker rats underwent treadmill exercise program. Body composition, strength, size, capillarization, insulin resistance (homeostatic model assessment [HOMA‐IR]) were measured. Apelin levels skeletal plasma then analyzed. Results Exercise improved body...
An experimental study was performed to investigate the axial-flexural load-carrying capacity of concrete-filled steel tube columns encased with thin precast concrete (PC). Six eccentrically loaded and one concentrically column were tested. To prevent premature failure encasement, various reinforcement details such as studs, fiber, welded wire mesh, cross ties used. The maximum axial loads specimens agreed strengths predicted by current design codes, although in some specimens, decreased...
Abstract Deep-tissue optical imaging suffers from the reduction of resolving power due to tissue-induced aberrations and multiple scattering noise. Reflection matrix approaches recording maps backscattered waves for all possible orthogonal input channels have provided formidable solutions removing severe recovering ideal diffraction-limited spatial resolution without relying on fluorescence labeling guide stars. However, measuring full input–output response tissue specimen is time-consuming,...
Composite columns of various sectional configurations are studied to maximize the contribution high-strength steel under compression-flexural loading. For concrete-encased (CES) columns, angles were placed at four corners cross-section and connected by types transverse reinforcements. In case a concrete-filled tube (CFT) column, slender was stiffened welding longitudinal bars. To verify structural capacities, eccentric axial load tests performed for three angle (CES-A) (S-CFT) using Grade...
Background We compared the predictive ability of our proposed N classification with that American Joint Committee on Cancer (AJCC) nodal (N) for oral cavity squamous cell carcinoma (OCC). Methods assessed 345 OCC patients who underwent primary tumor extirpation and neck lymph node (LN) dissection. was analyzed by recursive partitioning analysis AJCC c‐index. Cox proportional hazards regression analyses were used to determine associations between or factors disease‐free survival (DFS) overall...
Existing experimental results show that carefully designed RC column-to-steel beam (RCS) joints have high seismic resistance comparable to those of traditional steel or frames with details. However, such performance requires complicated and expensive strengthening details, which causes difficulties in fabrication construction. This study investigated simpler methods, targeting moderate capacity. Four interior RCS joint specimens were tested under cyclic lateral loading. The major test...
The development of deep learning has achieved great success in object detection, but small detection is still a difficult and challenging task computer vision. To address the problem, we propose an improved single-shot multibox detector (SSD) using enhanced feature map blocks (SSD-EMB). block (EMB) consists attention stream concatenation stream. allows proposed model to focus on regions rather than background owing channel averaging effectiveness normalization. provides additional semantic...
Many pathogenic bacteria form biofilms to survive under environmental stresses and host immune defenses. Differential expression (DE) analysis of the genes in biofilm planktonic cells a single condition, however, has limitations identify essential for formation. Independent component (ICA), machine learning algorithm, was adopted comprehensively Vibrio vulnificus, fulminating human pathogen, this study. ICA analyzed large-scale transcriptome data V. vulnificus various conditions then...
In the present study, a new fabrication method is proposed for concrete-filled hollow precast concrete columns (HPCC). method, corrugated steel plates were used internal mold of core section to improve production efficiency and structural safety. To evaluate constructability performance, mock-up tests cyclic loading performed four HPCC specimens conventional RC specimen. The test results showed that overall behavior was similar monolithic specimen, performances such as maximum load, initial...
This article describes the design, implementation, and evaluation of an active lane change control algorithm for autonomous vehicles with human factor considerations. Lane changes need to be performed considering both driver acceptance safety surrounding vehicles. Therefore, driving systems designed based on analysis behavior. In this article, manual characteristics are investigated using real-world test data. situations, interactions were mainly investigated. And indices developed kinematic...
Abstract Background This study compared our proposed N classification including the addition of number positive lymph node (LN+) and extranodal extension (ENE) with current previous American Joint Committee on Cancer (AJCC) classifications in salivary gland cancer (SGC). Methods involved 172 SGC patients who underwent surgery. Cox proportional hazard regression analyses were used to identify factors associated overall survival (OS) disease‐free (DFS). Results In multivariate analyses, LN+...
Abstract In the case of a visually impaired person, literal communication often relies on braille, system predominantly dependent vision and touch. This study entailed development visual tactile perception technique for braille character recognition. approach, recognition was performed using deep learning model (Faster R-CNN–FPN–ResNet-50), based custom-made dataset collected through data augmentation preprocessing. The attained performance indicated by an mAP50 94.8 mAP75 70.4 generated...
This paper describes design, vehicle implementation and validation of a motion planning control algorithm autonomous driving for lane change. Autonomous change is necessary high-level driving. A equipped with diverse devices like sensors computer introduced The system consists three parts: perception, control. In perception part, surrounding vehicles' states information are estimated. using these chassis information, probabilistic prediction conducted ego separately. And then, mode decided...
Colon transit time (CTT) is a useful diagnostic tool in chronic constipation, but requires good patient compliance. We analyzed the correlation between gas volume score (GVS) and CTT patients with constipation.The study included 145 consecutive (65 men) constipation. The primary outcome was be-tween colon GVS CTT. Secondary outcomes were differences according to subtypes of constipation.There 81 "CTT < 45 hours" 64 ≥ hours." In addition, 88 classi-fied as having functional constipation 57...