https://scholar.dgist.ac.kr/handle/20.500.11750/1194 2026-04-04T08:16:39Z https://scholar.dgist.ac.kr/handle/20.500.11750/60110 Title: SPATS: a practical system for comparative analysis of spatio-temporal graph neural networks Author(s): Yoon, Heeyong; Chon, Kang-Wook; Kim, Min-Soo Abstract: Thanks to technological advances in sensors and artificial intelligence, large amounts of data that combine spatial and temporal information are being produced in multiple domains. Spatio-temporal graph neural networks (STGNNs) have been recognized as highly effective models for analyzing spatio-temporal data, and so numerous novel STGNN models have recently been developed. However, no systematic and in-depth study has been carried out on the existing STGNN models with various datasets. Thus, it remains to be undecided whether more recent methods achieve better performance than traditional approaches. In this study, we propose a practical system, called SPAtio-Temporal graph System (SPATS), that performs effectively and efficiently the fair comparison of various STGNN models and datasets. SPATS introduces a unified data format to reduce dependency on data models and exploits GPU clusters to handle a large number of model comparisons automatically. Extensive experiments demonstrate that SPATS can efficiently compare STGNN models with reduced memory footprints and fully exploit GPU clusters. Furthermore, SPATS allows us to easily find the effective combination between the STGNN models and the datasets in various domains that have not been examined before. 2025-08-31T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/60108 Title: Neuroprotective and Neurotrophic Potential of Flammulina velutipes Extracts in Primary Hippocampal Neuronal Culture Author(s): Mitra, Sarmistha; Dash, Raju; Bashar, Md Abul; Mazumder, Kishor; Moon, Il Soo Abstract: Flammulina velutipes (enoki mushroom) is a functional edible mushroom rich in antioxidants, polysaccharides, mycosterols, fiber, and minerals. Accumulating evidence highlights its therapeutic potential across diverse pathological contexts, including boosting cognitive function. However, its role in neuromodulation has not been systematically explored. This study examined the effects of methanolic and ethanolic extracts of F. velutipes on primary hippocampal neurons. Neurons were treated with different extract concentrations, followed by assessments of cell viability, cytoarchitecture, neuritogenesis, maturation, and neuroprotection under oxidative stress. The extracts were further characterized by GC-MS to identify bioactive metabolites, and molecular docking combined with MM-GBSA binding energy analysis was employed to predict potential modulators. Our results demonstrated that the methanolic extract significantly enhanced neurite outgrowth, improved neuronal cytoarchitecture, and promoted survival under oxidative stress, whereas the ethanolic extract produced moderate effects. Mechanistic studies indicated that these neuroprotective and neurodevelopmental benefits were mediated through activation of the NTRK receptors, as validated by both in vitro assays and molecular docking studies. Collectively, these findings suggest that F. velutipes extracts, particularly methanolic fractions, may serve as promising neuromodulatory agents for promoting neuronal development and protecting neurons from oxidative stress. 2025-09-30T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/60107 Title: Subjective and Objective Quality Evaluation of Super-Resolution Enhanced Broadcast Images on a Novel SR-IQA Dataset Author(s): Kim, Yongrok; Shin, Junha; Lee, Juhyun; Ko, Hyunsuk Abstract: Super-Resolution (SR) is essential for displaying low-quality broadcast content on high-resolution screens. Recently, SR methods have been developed that not only increase resolution while preserving the original image information but also enhance the perceived quality. However, evaluating the quality of SR images generated from low-quality sources, such as SR-enhanced broadcast content, is challenging due to the need to consider both distortions and improvements. Additionally, assessing SR image quality without original high-quality sources presents another significant challenge. Unfortunately, there has been a dearth of research specifically addressing the Image Quality Assessment (IQA) of SR images under these conditions. In this work, we introduce a new IQA dataset for SR broadcast images in both 2K and 4K resolutions. We conducted a subjective quality evaluation to obtain Mean Opinion Score (MOS) for these SR images and performed a comprehensive human study to identify key factors influencing perceived quality. Finally, we evaluated the performance of existing IQA metrics on our dataset. This study reveals the limitations of current metrics, highlighting the need for a more robust IQA metric that better correlates with the perceived quality of SR images. The proposed dataset and the subjective evaluation platform are publicly available at https://sites.google.com/hanyang.ac.kr/ivml/datasets/sreb 2025-08-31T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/59959 Title: Real-Time Inductance Estimation of Sensorless PMSM Drive System Using Wavelet Denoising and Least-Order Observer with Time-Delay Compensation Author(s): Park, GwangMin; Bae, Junhyung Abstract: In this paper, the inductance of a sensorless PMSM (Permanent Magnet Synchronous Motor) drive system equipped with a periodic load torque compensator based on a wavelet denoising and least-order observer with time-delay compensation is estimated in real-time. In a sensorless PMSM system with constant load torque, the magnetically saturated inductance value remains constant. This constant inductance error causes minor performance degradation, such as a constant rotor position estimation error and non-optimal torque current, but it does not introduce a speed estimation error. Conversely, in a sensorless PMSM motor system subjected to periodic load torque, the magnetically saturated inductance error fluctuates periodically. This fluctuation leads to periodic variations in both the estimated position error and the speed error, ultimately degrading the load torque compensation performance. This paper applies the maximum energy-to-Shannon entropy criterion for the optimal selection of the mother wavelet in the wavelet transform to remove the motor signal noise and achieve more accurate inductance estimation. Additionally, the coherence and correlation theory is proposed to address the time delay in the least-order observer and improve the time delay. A self-saturation compensation method is also proposed to minimize periodic speed fluctuations and improve control accuracy through inductance parameter estimation. Finally, experiments were conducted on a sensorless PMSM drive system to verify the inductance estimation performance and validate the effectiveness of vibration reduction. 2025-11-30T15:00:00Z