https://scholarworks.bwise.kr/sch/handle/2021.sw.sch/806 2026-04-04T01:35:50Z https://scholarworks.bwise.kr/sch/handle/2021.sw.sch/28049 Title: Analysis of Near-Fall Detection Method Utilizing Dynamic Motion Images and Transfer Learning Authors: Kim, Jung-Yeon; Mat, Nab; Kim, Chomyong; Khan, Awais; Gil, Hyo-Wook; Lyu, Jiwon; Chung, Euyhyun; Kim, Kwang Seock; Jeon, Seob; Nam, Yunyoung Abstract: This study explores a model for detecting fall, non-fall and near-fall events as frequent experiences of near-falls are closely associated with a heightened risk of falls. Detecting near-falls can lead to more accurate predictions of falls. However, near-falls exhibit certain movement patterns similar to actual falls, making it challenging to distinguish between near-fall events and falls. We investigated the detection of fall-related activities, including falls, near-falls, and non-falls, by utilizing dynamic motion images derived from video clips. There were two primary classification approaches: a vanilla convolutional neural network (CNN) model and a transfer learning approach that utilizes InceptionV3 and DenseNet201 models as feature extractors and train conventional machine learning classifiers, such as support vector machine (SVM), K-nearest neighborhood, decision tree, and random forest, and adaptive boosting models. The vanilla CNN model achieved a high accuracy of 97.89% compared to the transfer learning approach, which reached a maximum accuracy of 95.54% for binary classification of fall and non-fall events. On the other hand, the transfer learning approach, which integrated feature from InceptionV3 and DenseNet201 into machine learning classifiers, achieved an accuracy of up to 90.14% for the three-class classification of fall, non-fall, and near-fall events. The findings of this study underscores the model & Atilde;s robustness in detecting various fall-related activities, highlighting its potential for improving safety in at-risk populations. 2025-12-01T00:00:00Z https://scholarworks.bwise.kr/sch/handle/2021.sw.sch/28070 Title: Increased risk of dementia in patients with primary Sjogren's syndrome: a nationwide population-based cohort study Authors: Lee, Kyung-Ann; Jeon, Hyeji; Kim, Hyun-Sook; Choi, Kyomin; Seo, Gi Hyeon Abstract: Background/Aims: This nationwide cohort study aimed to evaluate (1) whether primary Sjogren's syndrome (pSS) can contribute to the development of dementia and (2) whether the use of hydroxychloroquine (HCQ) can decrease the incidence of dementia in patients with pSS using the Health Insurance Review and Assessment database. Methods: We established a cohort between 2008 and 2020 of 20,160 patients with pSS without a history of dementia. The control group comprised sex- and age-matched individuals with no history of autoimmune disease or dementia. Cox proportional hazard analyses were performed to identify the association between pSS and dementia development. We also assessed the hazard ratio (HR) of dementia in early users of HCQ (within 180 days of the diagnosis of pSS) compared to non-users, adjusted for age, sex, and comorbidities. Results: The incidence of dementia was 0.68 (95% CI 0.64-0.72) cases per 100 person-years in pSS, and it was 0.58 (0.56-0.60) in the controls. The adjusted HR (aHR) of developing dementia was 1.16 (1.09-1.25) times greater in the pSS group than in the controls. The risk of dementia did not increase in HCQ users (aHR 1.07 [0.94-1.21]), but HCQ non-users had a 1.22 (1.12-1.33) higher risk of developing dementia than the matched controls. The use of HCQ lowered the risk of dementia in comparison with non-users in patients with pSS (aHR 0.82 [0.71-0.94]). Conclusions: Our results suggest that pSS is associated with an increased risk of dementia. HCQ may prevent dementia in patients with pSS. 2025-03-01T00:00:00Z https://scholarworks.bwise.kr/sch/handle/2021.sw.sch/28175 Title: Recycling Motorcycle Exhaust Soot into Fluorescent Graphene Oxide Quantum Dots for Sensing Ferrocyanide Ions and Bioimaging Cells: A Method for Waste Utilization Authors: Das, Chanchal; Sepay, Nasim; Kim, Tae Wan; Chae, Shinwon; Ghosh, Nandan; Dumpala, Mohan; Choi, Dongsic; Jeon, Seob; Im, Jungkyun; Biswas, Goutam Abstract: Graphene oxide quantum dots (GOQDs) with a high quantum yield (50%) were synthesized using soot collected from a motorcycle (petroleum vehicle) exhaust pipe and applied as sensors for ferrocyanide ([Fe(CN)6]4-) ions and as bioimaging agents in a cancer cell line. X-ray photoelectron spectroscopy (XPS) data for the GOQDs revealed a C/O ratio of 2.49, which was close to that of graphene oxide (GO). The synthesized GOQDs exhibited strong blue fluorescence. High sensitivity to detect [Fe(CN)6]4- was reported in GOQDs with a detection limit of 0.46 nmol mL-1, and a strong linear relationship was achieved in the concentration range of 100-1100 mu g L-1. The results demonstrate the utility of GOQDs for detecting [Fe(CN)6]4- in a real scenario. The GOQDs exhibited almost negligible cytotoxicity in cells and were internalized within 4 h of incubation, emitting blue fluorescence in the cytoplasm. This suggests that the GOQDs are promising bioimaging agents for biomedical applications. In general, these waste-derived GOQDs appear to be good chemo- and biosensing probes for real-life applications. 2025-03-01T00:00:00Z https://scholarworks.bwise.kr/sch/handle/2021.sw.sch/27670 Title: Ensemble Deep Learning for Classifying Sleeping Posture of Humans Covered in Blankets Using RGB and Thermal Imaging Authors: Khan, Awais; Kim, Chomyong; Mey, Senghour; Kim, Kwang Seock; Chung, Euyhyun; Lyu, Jiwon; Gil, Hyo-Wook; Jeon, Seob; Nam, Yunyoung Abstract: Accurate sleep posture monitoring plays a pivotal role in the diagnosis and treatment of a spectrum of sleep-related disorders, such as sleep apnea, restless leg syndrome, and rapid eye movement sleep behavior disorder. These disorders can significantly affect an individual's overall well-being and quality of life. However, the complexity of the human body and various factors affecting sleep patterns pose substantial challenges. Thus, this study introduces a sleep position determination approach that utilizes both RGB and thermal cameras to acquire visual and thermal data. This dual-source data acquisition system enables a comprehensive analysis of body positioning during sleep, including potential discomfort assessment. The methodology for this endeavor was initiated by acquiring a dataset encompassing various sleep postures, which was achieved through the deployment of RGB and thermal cameras. This dataset includes video footage of five frequently adopted sleep positions: supine, left log, right log, prone left, and prone right, involving the participation of nine individuals. Furthermore, the dataset was collected under two conditions: with and without a blanket. The proposed method begins by normalizing the database to the video frames. Next, the fine-tuned MobileNet-V2 and Inception-V3 models were employed for feature extraction. The tree-seed algorithm was used to select optimal features from the extracted data, reduce dimensionality, and improve the classification performance. Subsequently, Parallel Standard Deviation Padding Max Value (PSPMV), was applied to combine the feature vectors from the RGB and thermal datasets to enhance the accuracy. The fused vectors are then classified using ensemble machine learning models. Our method achieved an accuracy of 97.8% and 98.4% with and without blanket, using subspace KNN. After applying PSPMV fusion to blanket features, the accuracy improved to 99.2%. 2024-12-01T00:00:00Z