연구성과로 돌아가기
2023 연구성과 (269 / 285)
※ 컨트롤 + 클릭으로 열별 다중 정렬 가능합니다.
Excel 다운로드
| WoS | SCOPUS | Document Type | Document Title | Abstract | Authors | Affiliation | ResearcherID (WoS) | AuthorsID (SCOPUS) | Author Email(s) | Journal Name | JCR Abbreviation | ISSN | eISSN | Volume | Issue | WoS Edition | WoS Category | JCR Year | IF | JCR (%) | FWCI | FWCI Update Date | WoS Citation | SCOPUS Citation | Keywords (WoS) | KeywordsPlus (WoS) | Keywords (SCOPUS) | KeywordsPlus (SCOPUS) | Language | Publication Stage | Publication Year | Publication Date | DOI | JCR Link | DOI Link | WOS Link | SCOPUS Link |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ○ | Conference paper | MicroVM on Edge: Is It Ready for Prime Time? | Container virtualization is recognized as indispensable for realizing the vision of edge computing due to its advantages. However, OS-level virtualization suffers from a relatively low degree of security. Recently, microVM technology has emerged in response to this deficiency to provide stronger isolation and security while delivering performance comparable to the containers. In this work, we aim to gain a better understanding of microVM's suitability for edge computing in comparison with containers. We conduct extensive experiments on diverse workloads to test how microVMs compare against containers in several aspects. Through rigorous measurements and analysis, we extract several important findings. Despite having a more complex architecture than containers, microVMs perform comparably to the containers in terms of I/o performance. MicroVMs can even outperform containers in certain I/O workload types by 69%. Network I/O performance of microVMs can be 3x better than containers. We provide our findings and insights on the performance characteristics of microVMs on edge. © 2023 IEEE. | Lee, Kyungwoon; Tak, Byungchul | Kyungpook National University, Daegu, South Korea; Kyungpook National University, Daegu, South Korea | 57190025432; 6506911621 | bctak@knu.ac.kr; | Proceedings - IEEE Computer Society's Annual International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunications Systems, MASCOTS | 1526-7539 | 0.52 | 2025-06-25 | 1 | Edge; Firecracker; Kata Containers; MicroVm | Edge computing; Virtual reality; Virtualization; Edge; Edge computing; Firecracker; Kata container; Low degree; Measurement and analysis; Microvm; Performance; Prime time; Virtualizations; Containers | English | Final | 2023 | 10.1109/mascots59514.2023.10387638 | 바로가기 | 바로가기 | |||||||||||||||||
| ○ | Article | Mineral content and antioxidant potential of Pu-erh tea of different storage periods | Pu-erh tea is a health-promoting beverage that is consumed in many countries; however, the nutritional value of the tea is significantly affected by various factors including processing and storage. The objective of this study was to investigate the mineral contents and antioxidant potentials of 2-, 9-, and 21-year-old Pu-erh teas. The results demonstrated that the total mineral content of fermented tea (FT) was higher than those of raw tea (RT) after 9 years and 21 years. The 9-year-old FT had the highest K content, while the 2-year-old FT had the lowest K content among the FT samples. In contrast, the 2-year-old RT had the highest K content, while that of the 21-year-old RT was the lowest among the RT samples. Interestingly, the 9-year-old RT had the highest Na, Ca, Mg, and Cu contents among the all of the sampels . Overall, these results reveal that Pu-erh tea stored for 9 years would be the most beneficial with regards to the mineral content and antioxidant potential. Furthermore, this study provides valuable insights into the changes that occur in the nutrient content of Pu-erh tea during storage. ©The Korean Society of Food Science and Technology. | Choi, Sung-Hee; Kim, Il-Doo; Kim, So-Hyun; Kwon, Jae-Il; Shin, Dong-Hyun | Department of Korean Culture, Wonkwang University, South Korea; International Institute of Research & Development, Kyungpook National University, South Korea; School of Applied Biosciences, Kyungpook National University, South Korea; Department of Travel·Airline Master, Yeungnam University College, South Korea; School of Applied Biosciences, Kyungpook National University, South Korea | 57204726906; 56269995600; 58090464600; 58416326400; 7403352903 | dhshin@knu.ac.kr; | Korean Journal of Food Science and Technology | 0367-6293 | 55 | 2 | 0 | 2025-06-25 | 0 | Antioxidant potential; extraction frequency; mineral content; Pu-erh tea; storage period | English | Final | 2023 | 10.9721/kjfst.2023.55.2.101 | 바로가기 | 바로가기 | ||||||||||||||||
| ○ | Book chapter | ML Algorithms and Their Approach on COVID-19 Data Analysis | This chapter begins with characterizing Supervised Learning and Unsupervised learning and investigates Machine Learning algorithms in every one of the sub domains of Regression, Classification, Clustering, and so forth. It also talks about the engineering of calculations like Linear Regression, Logistic Regression, K-Means, K Nearest Neighbors, Hierarchical, DB Scan, Decision Tree, Random Forest Regression, and Random Forest classifier. Utilization of every algorithm to investigate the dataset will be displayed by carrying out it on renowned dataset model, and output of each piece of code is displayed with their preview. This section likewise takes care of the issue of predicting the future number of COVID-19 cases and the precision behind each model or algorithm is shown and investigated utilizing different measurements dependent on situation or issue articulation, for example, either issue is on forecast or order. This chapter does not focus on the solution of COVID-19 data analysis or expectation, rather it will be followed and will task different models dependent on need with conclusive target being clear comprehension of the Machine Learning algorithms and its execution in Python. © 2023 Scrivener Publishing LLC. | Ashok, Kambaluru; Reddy, Penumalli Anvesh; Kumar, Kukatlapalli Pradeep | Department of CS, Kyungpook National University, Daegu, South Korea; Department of CSE, Christ University, Karnataka, Bangalore, India; Department of CSE, Christ University, Karnataka, Bangalore, India | 58646932200; 58647328700; 57208700874 | penumalli.anvesh@btech.christuniversity.in; | Data Engineering and Data Science: Concepts and Applications | 1.11 | 2025-06-25 | 1 | classification; covid analysis; covid data visualization; Machine learning; regression; supervised learning; un-supervised learning | English | Final | 2023 | 10.1002/9781119841999.ch14 | 바로가기 | 바로가기 | |||||||||||||||||||
| ○ | Article | Mobile Crane for Path Tracking of Humanoid Robots | This paper describes the design and control of a mobile crane used for walking experiments conducted on humanoid robots. To prevent humanoid robots from overturning, experiments are conducted while they are connected to a crane. To conduct the walking experiment, a minimum of one operator each is required for controlling the robot and moving the crane. Thus far, an automated mobile crane that can perform the function of a crane operator has not been developed. To obtain high stability and efficiency during the walking experiment, we propose a mobile crane that moves according to the locomotion of humanoid robots. Considering the locomotive property of humanoid robots, the mobile crane was designed to be able to move omnidirectionally, and a control algorithm was developed for the mobile crane to track the position and orientation of the humanoid robot being studied. The performance of the developed mobile crane was evaluated through a walking experiment conducted using a humanoid robot. © ICROS 2023. | Park, Shinwoo; Joe, Hyun-Min | Department of Robot and Smart System Engineering, Kyungpook National University, South Korea; Department of Robot and Smart System Engineering, Department of Artificial Intelligence, Kyungpook National University, South Korea | 58839299500; 57188687051 | hmjoe@knu.ac.kr; | Journal of Institute of Control, Robotics and Systems | 1976-5622 | 29 | 9 | 0.14 | 2025-06-25 | 1 | humanoid robot; mecanum wheel; mobile crane; omnidirectional mobile robot; ultrasonic sensor | Cranes; Machine design; Design and control; Humanoid robot; Mecanum wheels; Mobile cranes; Omnidirectional mobile robot; Path tracking; Performance; Position and orientations; Property; Anthropomorphic robots | Korean | Final | 2023 | 10.5302/j.icros.2023.23.0048 | 바로가기 | 바로가기 | |||||||||||||||
| ○ | Conference paper | Model Predictive Control of Five-Level Current Source Converter with Optimized Weighting Factors | This paper presents the finite control set model predictive control (FCS-MPC) with improved control objectives using optimized weighting factors. The Pareto curve obtained from the Particle Swarm optimizes the weighting factors. The optimal weighting factors are then used in the FCS-MPC algorithm to enhance the control objective performance of a three-phase five-level current source converter (5L-CSC). In 5L-CSC, two identical three-level CSC (3L-CSC) modules are connected in parallel with one common current source supply to provide uniform five-level output current waveform. This CSC requires a current balance to operate the converter safely. Therefore, the control objective of the MPC is to control the output voltage and balance the internal currents. The weighting factors are optimized based on the converter's total harmonics distortion (THD) performance and inductor current ripple. The performance enhancement of 5L-CSC, utilizing optimal weighting factors, is confirmed through MATLAB simulations when compared to the model employing arbitrarily assigned weighting factors. The simulation outcomes additionally indicate that the integration of Particle Swarm Optimization (PSO) with FCS-MPC exhibits resilience across various condition settings. © 2023 IEEE. | Chen, Muyu; Yuan, Zhige; Faraji, Faramarz; Ghias, Amer M. Y. M.; Cha, Honnyong; Nicholas, Vun Chan Hua | School of Electrical and Electronic Engineering, Nanyang Technological University, Collab. Initiative Interdisciplinary Graduate Programme, Singapore; School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore; School of Energy Engineering, Kyungpook National University, Daegu, South Korea; School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore; School of Energy Engineering, Kyungpook National University, Daegu, South Korea; School of Computer Science and Engineering, Nanyang Technological University, Singapore | 58650313900; 58651219800; 57191226987; 46062041900; 24450248400; 58650767900 | 2023 IEEE 3rd International Conference on Industrial Electronics for Sustainable Energy Systems, IESES 2023 | 0 | 2025-06-25 | 0 | Current Source Converter; Finite Control Set Model Predictive Control; Multi-Objective Particle Swarm Optimization; Weighting Factor Optimization | Curve fitting; MATLAB; Multiobjective optimization; Particle swarm optimization (PSO); Power converters; Predictive control systems; Control objectives; Current source converters; Factor optimization; Finite control set; Finite control set model predictive control; Model-predictive control; Multi objective particle swarm optimization; Set models; Weighting factor optimization; Weighting factors; Model predictive control | English | Final | 2023 | 10.1109/ieses53571.2023.10253745 | 바로가기 | 바로가기 | |||||||||||||||||||
| ○ | Conference paper | Modification of roof structure of the traditional wood houses in the urban area of daegu, korea during mid twentieth century | This study focuses on the roof structure modification of hanok, the traditional Korean timber frame house, in the central urban area of Daegu city, Korea during mid twentieth century. It is based on the 2 years survey of 10,753 cases of hanok architectures in this city. The urban sprawl and population growth gave an impetus to house builders developing new residential areas with modifing traditional plan and structure of hanok. Instead of traditional way, the wood house builders of this period choose simpler and cheaper roof structure for housing in the dense urban area. They wanted to make the familiar and popular image of tile-roofed traditional wood house using heavy timber frame with cheaper cost and shorter construction time for adapting to housing market. New method using sparsely arranged small square cross-section rafters was in fashion. Other roof type with short and asymmetric eaves was used at small sites. Some special houses have newly designed structure for long span, too. These modified wood houses shows that there were diverse needs at the housing market. But after this period wood house lost the popularity in the market as modern house building technology, and reinforced concrete construction was widely utilized. Copyright © (2023) WORLD CONFERENCE ON TIMBER ENGINEERING 2023 (WCTE 2023) All rights reserved. | Cho, Jaemo | Kyungpook National University, South Korea | 57189308661 | zozemo@gmail.com; | 13th World Conference on Timber Engineering, WCTE 2023 | 7 | 0 | 2025-06-25 | 0 | Asymmetric eave; Housing market; Korean timber frame house hanok; Roof structure; Thin rafter | Commerce; Concrete construction; Population statistics; Reinforced concrete; Roofs; Timber; Urban growth; Wooden buildings; Wooden construction; Asymmetric eave; House builders; Housing markets; Korean timber frame house hanok; Roof structures; Thin rafte; Timber frame house; Twentieth century; Urban areas; Wood house; Houses | English | Final | 2023 | 10.52202/069179-0533 | 바로가기 | 바로가기 | |||||||||||||||||
| ○ | Article | Molecular and Morphological Characterization of Three New Fungal Species of the Class Sordariomycetes from Korea | Three fungal strains belonging to the class Sordariomycetes were isolated from soil collected on Jeju Island and Gyeongsangbuk-do, Korea. They were identified as Diaporthe endophytica (KNU-JJ-1809), Faurelina indica (KNU-JJ-1830), and Trichoderma ivoriense (KNU-4-KH1). KNU-JJ-1809 produced beta conidia that were straight, curved, hyaline, smooth-walled, with a diameter of 16.5-25.0×0.6-1.7 μm. The conidia of strain KNU-JJ-1830 were hyaline to light green, thin, clavate, round, truncate base, had guttules at both ends, with a diameter of 2.5-5.2×1.7-3.8 μm. The conidia of strain KNU-4-KH1 were oblong or ellipsoidal, smooth-walled, greenish, with a diameter of 2.2-4.4×2.2-3.6 μm. Internal transcribed spacer regions, partial large subunit, translation elongation factor 1-alpha, β-tubulin, and calmodulin genes were used to confirm the strains, and their cultural and morphological characteristics. To our knowledge, this is the first report on D. endophytica, F. indica, and T. ivoriense in Korea. © 2023 THE KOREAN SOCIETY OF MYCOLOGY. | Das, Kallol; Ban, Jae-Ho; Choi, So-Young; Lee, Seung-Yeol; Jung, Hee-Young | College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, South Korea; College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, South Korea, Plant Quarantine Technology Center, Animal and Plant Quarantine Agency, Gimcheon, 39660, South Korea; College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, South Korea; College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, South Korea, Institute of Plant Medicine, Kyungpook National University, Daegu, 41566, South Korea; College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, South Korea, Institute of Plant Medicine, Kyungpook National University, Daegu, 41566, South Korea | 57203751520; 57244356500; 57202918688; 56106499600; 7403029383 | leesy1123@knu.ac.kr; | Korean Journal of Mycology | 0253-651X | 51 | 1 | 0 | 2025-06-25 | 0 | Diaporthe endophytica; Faurelina indica; Sordariomycetes; Trichoderma ivoriense | English | Final | 2023 | 10.4489/kjm.20230003 | 바로가기 | 바로가기 | ||||||||||||||||
| ○ | ○ | Article | Molecular epidemiology of Toxoplasma gondii in cattle in Korea | Toxoplasmosis is a major public health concern, with raw or undercooked meat being the primary source of human infection. Knowledge regarding the molecular epidemiolo-gy of Toxoplasma gondii in cattle destined for human consumption in Korea is lacking. The present study aimed to genetically characterize the infectious strains of T. gondii. Overall, 455 cattle blood samples from 84 farms in the Gyeongnam-do (Province) were randomly collected in 2017. Nested PCR analysis revealed that only 3 (0.7%) samples were infected with T. gondii. The B1 gene sequence of T. gondii was observed to be similar (97.3-99.6%) to that of other T. gondii isolates. This is the first study to perform the molecular detection of T. gondii in cattle in Korea. Although the prevalence of infec-tion was low, our findings suggest that cattle present a potential public health issue. It may be crucial to recognize the importance of T. gondii infection in cattle meat. | Kim, Kyoo-Tae; Seo, Min -Goo | Kyungpook Natl Univ, Coll Vet Med, Daegu 41566, South Korea | ; Seo, Min-Goo/NQF-4335-2025 | 56680415000; 53982155300 | koreasmg@knu.ac.kr; | PARASITES HOSTS AND DISEASES | PARASITE HOST DIS | 2982-5164 | 2982-6799 | 61 | 2 | SCIE | PARASITOLOGY | 2023 | N/A | 0.71 | 2025-06-25 | 3 | 3 | Toxoplasma gondii; cattle; phylogenetic analysis; genotype; prevalence | IDENTIFICATION; ANTIBODIES; INFECTION; ANIMALS | cattle; genotype; phylogenetic analysis; prevalence; Toxoplasma gondii | Animals; Cattle; Humans; Meat; Molecular Epidemiology; Republic of Korea; Toxoplasma; Toxoplasmosis, Animal; genomic DNA; animal experiment; Article; bacterium isolate; blood sampling; bovine tuberculosis; cattle breed; dairy cattle; DNA extraction; enzyme linked immunosorbent assay; Escherichia coli; female; gene sequence; Korea; male; maximum likelihood method; molecular epidemiology; nested polymerase chain reaction; nonhuman; phylogeny; prevalence; sequence alignment; Toxoplasma gondii; animal; animal toxoplasmosis; bovine; epidemiology; genetics; human; meat; molecular epidemiology; South Korea; Toxoplasma | English | 2023 | 2023-05 | 10.3347/phd.23016 | 바로가기 | 바로가기 | 바로가기 | 바로가기 | ||
| ○ | Article | Morphological and Phylogenetic Analysis of a New Record ofParaconiothyrium kelleni from Soil in Korea | A fungal strain designated KNUF-21-66Q1 was isolated from soil in Chungcheongbuk Province, Korea. Moderate growth of colonies was observed on potato dextrose agar, oatmeal agar (OA), malt extract agar, and cornmeal agar media at 25℃, and the detailed morphology was examined on OA medium. The colonies on OA medium were flat, had entire margin, hyaline, and yellow concentric rings in 3-4 weeks. Conidiomata were pycnidial, solitary or clustered, globose to subglobose, black-brown, and 300-500 µm in diameter. Conidiogenous cells were smooth, hyaline, globose to ampulliform, and 6.0-9.0×3.0-6.0 µm in size (n=15). Conidia were hyaline to pale brown, slightly golden, obovoid to slightly ellipsoidal, smooth, guttulate, and 3.0-4.7×2.1-3.3 µm in size (n=100). The strain was confirmed based on phylogenetic analysis using internal transcribed spacer regions, the partial 28S rDNA of large subunit, and β-tubulin gene sequences. The morphological observations and phylogenetic analysis revealed that the strain KNUF-21-66Q1 was similar to the previously described Paraconiothyrium kelleni. To our knowledge, this is the first report of P. kelleni in Korea. © 2023 THE KOREAN SOCIETY OF MYCOLOGY. | Yadav, Mukesh Kumar; Das, Kallol; Ryu, Jung-Joo; Lim, Seong-Keun; Choi, Jin-Sil; Lee, Seung-Yeol; Jung, Hee-Young | College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, South Korea; College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, South Korea; College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, South Korea; College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, South Korea; College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, South Korea; College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, South Korea, Institute of Plant Medicine, Kyungpook National University, Daegu, 41566, South Korea; College of Agriculture and Life Sciences, Kyungpook National University, Daegu, 41566, South Korea, Institute of Plant Medicine, Kyungpook National University, Daegu, 41566, South Korea | 58512402600; 57203751520; 57226333274; 57756003900; 58512402700; 56106499600; 7403029383 | heeyoung@knu.ac.kr; | Korean Journal of Mycology | 0253-651X | 51 | 2 | 0 | 2025-06-25 | 1 | Morphology; Paraconiothyrium kelleni; Phylogeny; Soil-inhabiting fungi | English | Final | 2023 | 10.4489/kjm.20230011 | 바로가기 | 바로가기 | ||||||||||||||||
| ○ | ○ | Article | Mortality Prediction in Patients With or Without Heart Failure Using a Machine Learning Model | BACKGROUND Most risk prediction models are confined to specific medical conditions, thus limiting their application to general medical populations. OBJECTIVES The MARKER-HF (Machine learning Assessment of RisK and EaRly mortality in Heart Failure) risk model was developed in heart failure (HF) patients. We assessed the ability of MARKER-HF to predict 1-year mortality in a large community-based hospital registry database including patients with and without HF. METHODS This study included 41,749 consecutive patients who underwent echocardiography in a tertiary referral hospital (4,640 patients with and 37,109 without HF). Patients without HF were further subdivided into those with (n = 22,946) and without cardiovascular disease (n = 14,163) and also into cohorts based on recent acute coronary syndrome or history of atrial fibrillation, chronic obstructive pulmonary disease, chronic kidney disease, diabetes mellitus, hypertension, or malignancy. RESULTS The median age of the 41,749 patients was 65 years, and 56.2% were male. The receiver operated area under the curves for MARKER-HF prediction of 1-year mortality of patients with HF was 0.729 (95% CI: 0.706-0.752) and for patients without HF was 0.770 (95% CI: 0.760-0.780). MARKER-HF prediction of mortality was consistent across subgroups with and without cardiovascular disease and in patients diagnosed with acute coronary syndrome, atrial fibrillation, chronic obstructive pulmonary disease, chronic kidney disease, diabetes mellitus, or hypertension. Patients with malignancy demonstrated higher mortality at a given MARKER-HF score than did patients in the other groups. CONCLUSIONS MARKER-HF predicts mortality for patients with HF as well as for patients suffering from a variety of diseases. (JACC Adv 2023;2:100554) (c) 2023 The Authors. Published by Elsevier on behalf of the American College of Cardiology Foundation. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/). | Jang, Se Yong; Park, Jin Joo; Adler, Eric; Eshraghian, Emily; Ahmad, Faraz S.; Campagnari, Claudio; Yagil, Avi; Greenberg, Barry | Univ Calif San Diego, Dept Cardiol, 9454 Med Ctr Dr, La Jolla, CA 92037 USA; Kyungpook Natl Univ, Sch Med, Dept Internal Med, Div Cardiol, Daegu, South Korea; Seoul Natl Univ, Cardiovasc Ctr, Dept Internal Med, Div Cardiol,Bundang Hosp, Seoul, South Korea; Northwestern Univ, Feinberg Sch Med, Dept Med, Div Cardiol, Chicago, IL USA; Northwestern Med, Bluhm Cardiovasc Inst Ctr Artificial Intelligence, Chicago, IL USA; Univ Calif Santa Barbara, Phys Dept, Santa Barbara, CA USA; Univ Calif San Diego, Phys Dept, San Diego, CA USA | 57207977889; 35799900000; 59283823400; 57212930588; 54392524600; 59886195500; 59434614500; 7201629440 | bgreenberg@health.ucsd.edu; | JACC-ADVANCES | 2772-963X | 2 | 7 | 1.63 | 2025-06-25 | 1 | 5 | heart failure; MARKER-HF; mortality; risk score | IN-HOSPITAL MORTALITY; RISK STRATIFICATION; CLASSIFICATION; VALIDATION; DERIVATION | heart failure; MARKER-HF; mortality; risk score | English | 2023 | 2023-09 | 10.1016/j.jacadv.2023.100554 | 바로가기 | 바로가기 | 바로가기 | |||||||||||
| ○ | Conference paper | MPNet: Multiscale predictions based on feature pyramid network for semantic segmentation | Semantic segmentation is a complex topic where they assign each pixel of an image with a corresponding class and demand accuracy at objective boundaries. The method plays a vital role in scene-understanding scenarios. For self-driving applications, the input source includes various types of objects such as trucks, people, or traffic signs. One receptive field is only effective in capturing a short range of sizes. Feature pyramid network (FPN) utilizes different fields of view to extract information from the input. The FPN approach obtains the spatial information from the high-resolution feature map and the semantic information from the lower scales. The final feature representation contains coarse and fine details, but it has some drawbacks. They burden the system with extensive computation and reduce the semantic information. In this paper, we devise an effective multiscale predictions network (MPNet) to address these issues. A multiscale pyramid of predictions effectively processes the prominent characteristics of each feature. A pair of adjacent features is combined together to predict the output separately. A lower-scale feature of each prediction is assigned as the contextual contributor, and the other provides coarser information. The contextual branch is passed through the atrous spatial pyramid pooling to improve performance. The segmentation scores are fused to obtain advantages from all predictions. The model is validated by a series of experiments on open data sets. We have achieved good results 76.5% mIoU at 50 FPS on Cityscapes and 43.9% mIoU on Mapillary Vistas. © 2023 IEEE. | Quyen, Van Toan; Kim, Min Young | Kyungpook National University, Electronic and Electrical Engineering, Daegu, South Korea; Kyungpook National University, Electronic and Electrical Engineering, Daegu, South Korea | 57215669249; 56739349100 | yersin@knu.ac.kr; | International Conference on Ubiquitous and Future Networks, ICUFN | 2165-8528 | 2023-July | 1.49 | 2025-06-25 | 4 | feature pyramid network; multiscale prediction; real-time application; Semantic segmentation | Computer vision; Forecasting; Open Data; Semantic Web; Semantics; Traffic signs; Feature pyramid; Feature pyramid network; Multiscale predictions; Prediction-based; Pyramid network; Real-time application; Scene understanding; Self drivings; Semantic segmentation; Semantics Information; Semantic Segmentation | English | Final | 2023 | 10.1109/icufn57995.2023.10199608 | 바로가기 | 바로가기 | ||||||||||||||||
| ○ | ○ | Proceedings Paper | MS2P: A TRUE MULTI-VIEW SATELLITE STEREO PIPELINE WITHOUT RECTIFICATION OF PUSH BROOM IMAGES | In this study, we propose a new and true multi-view stereo pipeline (MS2P) to reconstruct surface model from push broom satellite images. The proposed method utilizes all unrectified multi-view stereo image at the same time, thus it is a true multi-view stereo (MVS) method. Using the rational polynomial coefficient (RPC) of the images, the ground-to-image geometry between all 3D cost volumes and multi-view images are defined. Then EnSoft3D, a true multi-view stereo algorithm, is used to iteratively compute and refine all height maps. In addition, VoteValue in the original EnSoft3D is adjusted to facilitate satellite images for more accurate height map generation. Experimental results show that an MVS method in the computer vision community can be used to reconstruct height map and surface model from multi-view unrectified satellite images. Based on experimental results using DFC19 dataset, the MSE and RMSE are about 0.572m and 2.09m in average. | Seo, DongUk; Lee, HyoSeong; Park, Soon-Yong | Kyungpook Natl Univ, Sch Elect & Elect Engn, Daegu, South Korea; Sunchon Natl Univ, Dept Civil Engn, Sunchon, South Korea | Park, Soon-Yong/HGV-2374-2022 | 58266434300; 58738860700; 7501834063 | IGARSS 2023 - 2023 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM | 2153-6996 | 1.27 | 2025-06-25 | 2 | 2 | Multi-view stereo; 3D reconstruction; Push broom; Satellite images | 3D reconstruction; Multi-view stereo; Push broom; Satellite images | English | 2023 | 2023 | 10.1109/igarss52108.2023.10282994 | 바로가기 | 바로가기 | 바로가기 | ||||||||||||||
| ○ | ○ | Proceedings Paper | Multi-attributed Face Synthesis for One-Shot Deep Face Recognition | Nothing is more unique and crucial to an individual's identity than their face. With the rapid improvement in computational power and memory space and recent specializations in deep learning models, images are becoming more essential than ever for pattern recognition. Several deep face recognition models have recently been proposed to train deep networks on enormously big public datasets like MSCeleb-1M [8] and VG-GFace2 [5], successfully achieving sophisticated performance on mainstream applications. It is particularly challenging to gather an adequate dataset that allows strict command over the desired properties, such as hair color, skin tone, makeup, age alteration, etc. As a solution, we devised a one-shot face recognition system that utilizes synthetic data to recognize a face even if the facial attributes are altered. This work proposes and investigates the feasibility of creating a multi-attributed artificial face dataset from a one-shot image to train the deep face recognition model. This research seeks to demonstrate how the image synthesis capability of the deep learning methods can construct a face dataset with multiple critical attributes for a recognition process to enable and enhance efficient face recognition. In this study, the ideal deep learning features will be combined with a conventional one-shot learning framework. We did experiments for our proposed model on the LFW and multi-attributed synthetic data; these experiments highlighted some insights that can be helpful in the future for one-shot face recognition. | Shaheryar, Muhammad; Laishram, Lamyanba; Lee, Jong Taek; Jung, Soon Ki | Kyungpook Natl Univ, Sch Comp Sci & Engn, Daegu, South Korea | Shaheryar, Muhammad/NBW-9729-2025; Jung, Soon Ki/P-7687-2018 | 56132068000; 57219930647; 24341317500; 57226791905 | shaheryar@knu.ac.kr;yanbalaishram@knu.ac.kr;jongtaeklee@knu.ac.kr;skjung@knu.ac.kr; | FRONTIERS OF COMPUTER VISION, IW-FCV 2023 | 1865-0929 | 1865-0937 | 1857 | 0 | 2025-06-25 | 0 | 0 | Deep Learning; Computer Vision; One-Shot Face recognition; Siamese Networks; Image Classification | Computer Vision; Deep Learning; Image Classification; One-Shot Face recognition; Siamese Networks | Computer vision; Deep learning; Image classification; Image enhancement; Learning systems; Computational memory; Computational power; Deep learning; Face synthesis; Images classification; One-shot face recognition; Power spaces; Recognition models; Siamese network; Synthetic data; Face recognition | English | 2023 | 2023 | 10.1007/978-981-99-4914-4_1 | 바로가기 | 바로가기 | 바로가기 | ||||||||||
| ○ | ○ | Proceedings Paper | Multi-Jet Event classification with Convolutional neural network at Large Scale | We present an application of Scalable Deep Learning to analyze simulation data of the LHC proton-proton collisions at 13 TeV. We built a Deep Learning model based on the Convolutional Neural Network (CNN) which utilizes detector responses as two-dimensional images reflecting the geometry of the Compact Muon Solenoid (CMS) detector. The model discriminates signal events of the R-parity violating Supersymmetry (RPV SUSY) from the background events with multiple jets due to the inelastic QCD scattering (QCD multi-jets). With the CNN model, we obtained x1.85 efficiency and x1.2 expected significance with respect to the traditional cut-based method. We demonstrated the scalability of the model at a Large Scale with the High-Performance Computing (HPC) resources at the Korea Institute of Science and Technology Information (KISTI) up to 1024 nodes. | Kim, Jiwoong; Moon, Chang-Seong; Nam, Hokyeong; Goh, Junghwan; Bae, Dongsung; Yoo, Changhyun; Kim, Sungwon; Kim, Tongil; Yoo, Hwidong; Hwang, Soonwook; Cho, Kihyeon; Hahm, Jaegyoon; Myung, Hunjoo; Kim, Minsik; Hong, Taeyoung | Korea Inst Sci & Technol Informat, 245 Daehak Ro, Daejeon, South Korea; Kyungpook Natl Univ, Dept Phys, 80 Daehakro, Daegu 41566, South Korea; Kyung Hee Univ, Dept Phys, 26 Kyungheedae Ro, Seoul 02447, South Korea; Yonsei Univ, Dept Phys, 50 Yonsei Ro, Seoul 03722, South Korea | Goh, Junghwan/Q-3720-2016; Moon, Chang-Seong/J-3619-2014 | 57225018658; 56365007800; 57835954400; 56448093200; 57836198100; 57207260043; 55924035800; 57835713100; 35228252700; 55863562700; 57207798241; 7006723687; 16646547400; 59072717700; 57190377043 | jiwoong.kim@cern.ch; | 20TH INTERNATIONAL WORKSHOP ON ADVANCED COMPUTING AND ANALYSIS TECHNIQUES IN PHYSICS RESEARCH | 1742-6588 | 1742-6596 | 2438 | 1.51 | 2025-06-25 | 1 | 1 | Convolution; Deep learning; High energy physics; Solenoids; Convolutional neural network; Detector response; Events classification; Large-scales; Learning models; Model-based OPC; Multi-jets; Proton proton collisions; Simulation data; Two dimensional images; Neural network models | English | 2023 | 2023 | 10.1088/1742-6596/2438/1/012103 | 바로가기 | 바로가기 | 바로가기 | ||||||||||||
| ○ | ○ | Article | Multi-level alignment for few-shot temporal action localization | Temporal action localization (TAL), which aims to localize actions in long untrimmed videos, requires a large number of annotated training data. However, it is expensive to obtain segment level annotations for large-scale datasets. To overcome this challenge, a new few-shot learning method is proposed that localizes temporal actions for unseen classes with only a few training samples. In this study, a new multi-level encoder cosine-similarity alignment module is adopted that exploits the alignment of visual information at each temporal location. The proposed method arranges the video snippets that contain similar foreground action instances, and it captures the intra-class variations more implicitly. In addition, it incorporates cosine similarity in Transformer encoder layers that supports the self-attention mechanism. This emphasizes more on refined features at the higher encoder layers. Towards this objective, an episodic-based training scheme is adopted to learn the alignment of similar video snippets with a few training examples. At the test time, the learned context information is then adapted to novel classes. Experimental results show that the proposed method outperforms the state-of-the-art methods for few-shot temporal action localization with single and multiple action instances on the ActivityNet-1.3 dataset and achieves competitive results on the THUMOS-14 and HACS datasets. | Keisham, Kanchan; Jalali, Amin; Kim, Jonghong; Lee, Minho | Kyungpook Natl Univ, Grad Sch Artificial Intelligence, Daegu 41566, South Korea; Kyungpook Natl Univ, AI Inst Technol, KNU LG Elect Convergence Res Ctr, Daegu 41566, South Korea; Keimyung Univ, Dept Res, Dongsan Med Ctr, 402,Daegu Technopk Keimyung Univ Ctr 1035, Daegu 42601, South Korea | Lee, Min-Ho/ABE-5735-2021; Jalali, Amin/AAH-6921-2019 | 57140837500; 57022190400; 57022250500; 57191730119 | kanchankeisham@gmail.com;max.jalali@gmail.com;jonghong89@gmail.com;mholee@gmail.com; | INFORMATION SCIENCES | INFORM SCIENCES | 0020-0255 | 1872-6291 | 650 | SCIE | COMPUTER SCIENCE, INFORMATION SYSTEMS | 2023 | N/A | 0.89 | 2025-06-25 | 7 | 8 | Few-shot learning; Temporal action localization; Feature alignment; Cosine similarity | ATTENTION | Cosine similarity; Feature alignment; Few-shot learning; Temporal action localization | Learning systems; Signal encoding; Annotated training data; Cosine similarity; Feature alignment; Few-shot learning; Large-scale datasets; Learning methods; Localisation; Multilevels; Temporal action localization; Training sample; Large dataset | English | 2023 | 2023-12 | 10.1016/j.ins.2023.119618 | 바로가기 | 바로가기 | 바로가기 | 바로가기 |
페이지 이동: