• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.07a
• /
• pp.313-315
• /
• 2021

Timely and accurate diagnosis of lung diseases using Chest X-ray images has been gained much attention from the computer vision and medical imaging communities. Although previous studies have presented the capability of deep convolutional neural networks by achieving competitive binary classification results, their models were seemingly unreliable to effectively distinguish multiple disease groups using a large number of x-ray images. In this paper, we aim to build an advanced approach, so-called Ensemble Knowledge Distillation (EKD), to significantly boost the classification accuracies, compared to traditional KD methods by distilling knowledge from a cumbersome teacher model into an ensemble of lightweight student models with parallel branches trained with ground truth labels. Therefore, learning features at different branches of the student models could enable the network to learn diverse patterns and improve the qualify of final predictions through an ensemble learning solution. Although we observed that experiments on the well-established ChestX-ray14 dataset showed the classification improvements of traditional KD compared to the base transfer learning approach, the EKD performance would be expected to potentially enhance classification accuracy and model generalization, especially in situations of the imbalanced dataset and the interdependency of 14 weakly annotated thorax diseases.

• Nuclear Engineering and Technology
• /
• v.55 no.7
• /
• pp.2516-2533
• /
• 2023

Several practical methods for accelerating the depletion calculation in a GPU-based Monte Carlo (MC) code PRAGMA are presented including the multilevel spectral collapse method and the vectorized Chebyshev rational approximation method (CRAM). Since the generation of microscopic reaction rates for each nuclide needed for the construction of the depletion matrix of the Bateman equation requires either enormous memory access or tremendous physical memory, both of which are quite burdensome on GPUs, a new method called multilevel spectral collapse is proposed which combines two types of spectra to generate microscopic reaction rates: an ultrafine spectrum for an entire fuel pin and coarser spectra for each depletion region. Errors in reaction rates introduced by this method are mitigated by a hybrid usage of direct online reaction rate tallies for several important fissile nuclides. The linear system to appear in the solution process adopting the CRAM is solved by the Gauss-Seidel method which can be easily vectorized on GPUs. With the accelerated depletion methods, only about 10% of MC calculation time is consumed for depletion, so an accurate full core cycle depletion calculation for a commercial power reactor (BEAVRS) can be done in 16 h with 24 consumer-grade GPUs.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.35 no.2
• /
• pp.155-159
• /
• 2022

Smart livestock, which combines information and communication technology (ICT) with livestock, can be said to be an effective solution to existing livestock problems such as productivity improvement, odors, and diseases. So far, it has hardly been universalized; thus, it is necessary to develop automation devices to reduce labor by localizing automation devices to expand the distribution of ICT technology to farms, and to advance precise specifications and health management technology using biometric information. Weighing scales currently being used in livestock farms are to prevent the spread of diseases by diagnosis and preparation for AI and other diseases in advance, using information on the growing weight of duck breeding. However, accurate values cannot be obtained due to poor breeding conditions. In this paper, we developed a separate data transmission system kit for the weighing scale and placed the sensor on top of the weighing scale so that the sensor wire is not affected by pollutants or ducks on the floor. A display function was provided, and a method of receiving and analyzing the serial port data of the weighing device, and then transmitting them to the data collection server was implemented.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.05a
• /
• pp.244-247
• /
• 2022

In recent years, the logistics industry has greatly driven the world's economic development. Due to the frequent occurrence of logistics information leakage and forgery, it is necessary to find a solution that can accurately trace the logistics information and ensure the security and authenticity of the logistics information. The birth of blockchain technology has enabled the logistics industry to realize the development from quantitative change to qualitative change. The distributed storage idea, decentralization characteristics, immutable nature, complex encryption algorithm, and other technical characteristics of the blockchain technology make it have a wide range of application prospects in the logistics industry. The purpose of this paper is to apply blockchain technology to the whole chain of logistics information traceability, to indirectly store the corresponding data generated by the logistics circulation link on the blockchain, and to combine the researched consensus algorithm and searchable encryption algorithm to form A set of logistics information traceability system to achieve efficient and accurate traceability of logistics information.

• Journal of Positioning, Navigation, and Timing
• /
• v.12 no.4
• /
• pp.399-408
• /
• 2023

In this paper, we propose a solution that enables continuous indoor/outdoor positioning of smartphone users through the integration of Pedestrian Dead Reckoning (PDR) and GPS/WiFi signals. Considering that accurate step detection affects the accuracy of PDR, we propose a Deep Neural Network (DNN)-based technology to distinguish between walking and non-walking signals such as walking in place. Furthermore, in order to integrate PDR with GPS and WiFi signals, a technique is used to select a proper measurement by distinguishing between indoor/outdoor environments based on GPS Dilution of Precision (DOP) information. In addition, we propose a technology to adaptively change the measurement error covariance matrix by detecting measurement outliers that mainly occur in the indoor/outdoor transition section through a residual-based χ² test. It is verified through experiments on a testbed that these technologies significantly improve the performance of PDR and PDR/GPS/WiFi fingerprinting-based integrated pedestrian navigation.

• Journal of Positioning, Navigation, and Timing
• /
• v.12 no.3
• /
• pp.237-244
• /
• 2023

A signal generation simulator is an economical and useful solution in Global Navigation Satellite System (GNSS) receiver design and testing. A software-defined radio approach is widely used both in receivers and simulators, and its flexible structure to adopt to new signals is ideally suited to the testing of a receiver and signal processing algorithm in the signal design phase of a new satellite-based navigation system before the deployment of satellites in space. The generation of highly accurate delayed sampled codes is essential for generating signals in the simulator, where its sampling rate should be chosen to satisfy constraints such as Nyquist criteria and integer and non-commensurate properties in order not to cause any distortion of original signals. A high sampling rate increases the accuracy of code delay, but decreases the computational efficiency as well, and vice versa. Therefore, the selected sampling rate should be as low as possible while maintaining a certain level of code delay accuracy. This paper presents the lower limits of the sampling rate for GNSS signal generation simulators. In the simulation, two distinct code generation methods depending on the sampling position are evaluated in terms of accuracy versus computational efficiency to show the lower limit of the sampling rate for several GNSS signals.

• Korean Journal of Artificial Intelligence
• /
• v.12 no.1
• /
• pp.17-24
• /
• 2024

This paper introduces a novel approach to time-series estimation for energy load forecasting within Virtual Power Plant (VPP) systems, leveraging advanced artificial intelligence (AI) algorithms, namely Long Short-Term Memory (LSTM) and Seasonal Autoregressive Integrated Moving Average (SARIMA). Virtual power plants, which integrate diverse microgrids managed by Energy Management Systems (EMS), require precise forecasting techniques to balance energy supply and demand efficiently. The paper introduces a hybrid-method forecasting model combining a parametric-based statistical technique and an AI algorithm. The LSTM algorithm is particularly employed to discern pattern correlations over fixed intervals, crucial for predicting accurate future energy loads. SARIMA is applied to generate time-series forecasts, accounting for non-stationary and seasonal variations. The forecasting model incorporates a broad spectrum of distributed energy resources, including renewable energy sources and conventional power plants. Data spanning a decade, sourced from the Korea Power Exchange (KPX) Electrical Power Statistical Information System (EPSIS), were utilized to validate the model. The proposed hybrid LSTM-SARIMA model with parameter sets (1, 1, 1, 12) and (2, 1, 1, 12) demonstrated a high fidelity to the actual observed data. Thus, it is concluded that the optimized system notably surpasses traditional forecasting methods, indicating that this model offers a viable solution for EMS to enhance short-term load forecasting.

• Steel and Composite Structures
• /
• v.50 no.4
• /
• pp.375-382
• /
• 2024

The network theory studies interconnection between discrete objects to find about the behavior of a collection of objects. Also, nanomaterials are a collection of discrete atoms interconnected together to perform a specific task of mechanical or/and electrical type. Therefore, it is reasonable to use the network theory in the study of behavior of super-molecule in nano-scale. In the current study, we aim to examine vibrational behavior of spherical nanostructured composite with different geometrical and materials properties. In this regard, a specific shear deformation displacement theory, classical elasticity theory and analytical solution to find the natural frequency of the spherical nano-composite structure. The analytical results are validated by comparison to finite element (FE). Further, a detail comprehensive results of frequency variations are presented in terms of different parameters. It is revealed that the current methodology provides accurate results in comparison to FE results. On the other hand, different geometrical and weight fraction have influential role in determining frequency of the structure.

• International Journal of Internet, Broadcasting and Communication
• /
• v.16 no.2
• /
• pp.50-58
• /
• 2024

Body temperature plays an important role in medicine, some diseases are characterized by changes in human body temperature. Monitoring body temperature also allows doctors to monitor the effectiveness of medical treatments. Accurate body temperature measurement is key to detecting fevers, especially fevers related to infection with the SARS-CoV-2 virus that caused the recent Covid-19 pandemic in the world. The solution of measuring body temperature using a thermal camera is fast but has a high cost and is not suitable for some organizations with difficult economic conditions today. Use a medical thermometer to measure body temperature directly for a slow rate, making it easier to spread disease from person to person. In this paper, we propose a completely automatic body temperature measurement system that can adjust the height according to the person taking the measurement, has a measurement logging system and is monitored via the internet. Experimental results show that the proposed method has successfully created a fully automatic human body measurement system. Furthermore, this research also helps the school's scientists and students gain more knowledge and experience to apply Internet of Things technology in real life.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2023.11a
• /
• pp.15-16
• /
• 2023

In this paper, we aimed to develop a new method for diagnosing the depth of neutralization in architectural and civil engineering structures using the core drilling method, which combines the speed of drilling with the accuracy of core ringing. When compared to the drilling method, the core drilling method showed a lower measurement deviation of 1-2mm (7.6%) in confirming the depth of neutralization. This is believed to be a result of potential interference during the sample collection process in the drilling method, where the drill may pass through aggregates, leading to overestimation, as indicated in previous studies. The rapid evaluation of neutralization using the core drilling method serves as an alternative to address the issues associated with both drilling and core ringing methods in diagnosing the depth of neutralization. It offers a solution to the inaccuracy caused by coarse aggregates and the cumbersome post-processing steps required for neutralization diagnosis. Our proposed technique aims to provide an accurate and expedited diagnosis of neutralization depth without the need for additional processes.