• Journal of The Korean Association of Information Education
• /
• 제25권6호
• /
• pp.1015-1024
• /
• 2021

There are various educational programming environments in which students can train artificial intelligence (AI) using block-based programming languages, such as Entry, Machine Learning for Kids, and Teachable Machine. However, these programming environments are designed so that students can train AI through a separate menu, and then use the trained model in the code editor. These approaches have the advantage that students can check the training process more intuitively, but there is also the disadvantage that both the training menu and the code editor must be used. In this paper, we present a novel artificial intelligence block that can perform both AI training and programming in the code editor. While this AI block is presented as a Scratch block, the training process is performed through a Python server. We describe the blocks in detail through the process of training a model to classify a blue pen and a red pen, and a model to classify a dental mask and a KF94 mask. Also, we experimentally show that our approach is not significantly different from Teachable Machine in terms of performance.

• Journal of Korean Society of Water and Wastewater
• /
• 제36권4호
• /
• pp.239-248
• /
• 2022

The prediction of algal bloom is an important field of study in algal bloom management, and chlorophyll-a concentration(Chl-a) is commonly used to represent the status of algal bloom. In, recent years advanced machine learning algorithms are increasingly used for the prediction of algal bloom. In this study, XGBoost(XGB), an ensemble machine learning algorithm, was used to develop a model to predict Chl-a in a reservoir. The daily observation of water quality data and climate data was used for the training and testing of the model. In the first step of the study, the input variables were clustered into two groups(low and high value groups) based on the observed value of water temperature(TEMP), total organic carbon concentration(TOC), total nitrogen concentration(TN) and total phosphorus concentration(TP). For each of the four water quality items, two XGB models were developed using only the data in each clustered group(Model 1). The results were compared to the prediction of an XGB model developed by using the entire data before clustering(Model 2). The model performance was evaluated using three indices including root mean squared error-observation standard deviation ratio(RSR). The model performance was improved using Model 1 for TEMP, TN, TP as the RSR of each model was 0.503, 0.477 and 0.493, respectively, while the RSR of Model 2 was 0.521. On the other hand, Model 2 shows better performance than Model 1 for TOC, where the RSR was 0.532. Explainable artificial intelligence(XAI) is an ongoing field of research in machine learning study. Shapley value analysis, a novel XAI algorithm, was also used for the quantitative interpretation of the XGB model performance developed in this study.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• 제22권3호
• /
• pp.1-8
• /
• 2022

Recently, Google MedeiaPipe presents a novel approach for neural network-based hair segmentation from a single camera input specifically designed for real-time, mobile application. Though neural network related to hair segmentation is relatively small size, it produces a high-quality hair segmentation mask that is well suited for AR effects such as a realistic hair recoloring. However, it has undesirable segmentation effects according to hair styles or in case of containing noises and holes. In this study, the energy function of the test image is constructed according to the estimated prior distributions of hair location and hair color likelihood function. It is further optimized according to graph cuts algorithm and initial hair region is obtained. Finally, clustering algorithm and image post-processing techniques are applied to the initial hair region so that the final hair region can be segmented precisely. The proposed method is applied to MediaPipe hair segmentation pipeline.

• Proceedings of the Korean Society of Propulsion Engineers Conference
• /
• 한국추진공학회 2004년도 제22회 춘계학술대회논문집
• /
• pp.369-375
• /
• 2004

Gas turbine engine simulation in terms of transient, steady state performance and operational characteristics is complex work at the various engineering functions of aero engine manufacturers. Especially, efficiency of control system design and development in terms of cost, development period and technical relevance implies controlling diverse simulation and identification activities. The previous engine simulation has been accomplished within a limited analysis area such as fan, compressor, combustor, turbine, controller, etc. and this has resulted in improper engine performance and control characteristics because of limited interaction between analysis areas. In this paper, we propose a new simulation methodology for gas turbine engine performance analysis as well as its digital controller to solve difficulties as mentioned above. The novel method has particularities of (ⅰ) resulting in the integrated control simulation using almost every component/module analysis, (ⅱ) providing automated math model generation process of engine itself, various engine subsystems and control compensators/regulators, (ⅲ) presenting total sophisticated output results and easy understandable graphic display for a final user. We call this simulation system GT3GS (Gas Turbine 3D Graphic Simulator). GT3GS was built on both software and hardware technology for total simulation capable of high calculation flexibility as well as interface with real engine controller. All components in the simulator were implemented using COTS (Commercial Off the Shelf) modules. In addition, described here includes GT3GS main features and future works for better gas turbine engine simulation.

• Journal of Powder Materials
• /
• 제30권2호
• /
• pp.140-145
• /
• 2023

Although the Ti-6Al-4V alloy has been used in the aircraft industry owing to its excellent mechanical properties and low density, the low formability of the alloy hinders broadening its applications. Recently, laser-powder bed fusion (L-PBF) has become a novel process for overcoming the limitations of the alloy (i.e., low formability), owing to the high degree of design freedom for the geometry of products having outstanding performance used in high-tech applications. In this study, to investigate the effect of bulk shape on the microstructure and mechanical properties of L-PBFed Ti-6Al-4V alloys, two types of samples are fabricated using L-PBF: thick and thin samples. The thick sample exhibits lower strength and higher ductility than the thin sample owing to the larger grain size and lower residual dislocation density of the thick sample because of the heat input during the L-PBF process.

• Journal of the Korean Geotechnical Society
• /
• 제39권8호
• /
• pp.41-48
• /
• 2023

Microbially induced calcite precipitation(MICP) is a novel cementation method meant to enhance soil engineering properties through the use of microorganisms. This study investigates the effect of different MICP concentrations on plant growth. Tall fescue seeds are grown in plant columns filled with Jumunjin sand. Following plant growth, the soil samples are treated with MICP via spraying method. The results indicate that the MICP-treated plants exhibit hampered growth compared with the untreated plants. pH and electrical conductivity(EC) tests are performed to analyze the changes in soil properties by MICP. The MICP-treated soils exhibit a pH = 7, similar to the untreated soil. However, the EC dramatically increases with the increase in the MICP concentration, which leads to an increase in the osmotic pressure of the soil surrounding the plant roots. Eventually, the higher osmotic pressure in MICP-treated soil hinders the absorption of water and nutrients in plant roots, thus inhibiting plant growth.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 한국정보통신학회 2022년도 춘계학술대회
• /
• pp.55-57
• /
• 2022

In this study, a customized semi-passive RFID receiver module was implemented for in-VIVO deep tissue photo-therapy. A novel wireless technique is required due to a limitation of RF communication in body environment, as internal body has a complex structure such as, skin, fat, skeleton, water, and so on. Recently, coherently incoherent beamforming (CIB) based on RFID was introduced and it is able to transmit wireless signal with high reliability under the incoherent condition such as in-VIVO deep tissue. The proposed miniature photo capsule based on RFID consists of miniature controller, ultra small LED array and wireless RFID chip. RF Reader can access with standard RFID protocol (ISO 18000-6c) using UHF RFID antenna, a control command is wirelessly writtern on USER Bank memory. With received control command, therapy LED array dims with mulilevel under timer control. The signal process of designed RFID photo therapy capsule is analyzed and evaluated under the various environments in detailed.

• Journal of the Korean Society of Industry Convergence
• /
• 제26권3호
• /
• pp.471-481
• /
• 2023

Digital twin is considered as a key technology of industry 4.0, thus being essential for the future of industrial production. Despite the significance, a systematic analysis of its technological landscape is lacking. This study aims to investigate the technological development trends and newly emerging technological convergence opportunities in the domain of digital twin by exploiting patent information derived from U SPTO. For this purpose, this study visualized and predicted the convergence dynamics among patent classification codes by adopting patent co-classification analysis and link prediction approach. The findings show that the number of digital twin-related patent applications has increased significantly since 2018. The CPC code G06F showed the highest eigenvector centrality, while G05B was characterized by highest betweenness centrality. According to the predictive model, 41 novel links were revealed, acting as potential technology convergence opportunities. These links were then categorized into 11 different domains. The most dominant category was "digital data processing and artificial intelligence", which could play a foundational role in the diffusion of digital twin technology. The presence of digital twin technology is dominant in manufacturing, but its applications are expected to expand, including "climate change", "healthcare" and "aerospace engineering". The derived insights can support R&D managers and policy makers in formulating R&D strategies and directing future R&D investment decisions.

• Journal of the Korea Convergence Society
• /
• 제13권1호
• /
• pp.161-166
• /
• 2022

HIF-2α is a transcription factor activated mainly in hypoxic condition known to play crucial roles in a wide variety of pathophysiological events including cancer, metabolic syndrome, arthritis etc. In this context, a number of N'-aryl isonicotinolyhydrazides, in which known pharmacophores are included, have been selected from commercial chemical library and tested for the inhibitory activities targeting HIF-2α in cultured HTB94 cell. HRE-luciferase and HIF-2α were introduced into the cell by transfection and adenoviri infection, respectively and the reporter gene assay discovered the potency of 2-hydroxy-1-naphthyl structure. Accordingly, the scaffold has been adjusted based on this structure and subjected to anti-HIF-2α activity test, identifying 2 compounds as HIF-2α inhibitors. The activities were confirmed by false positive test. This study has been performed via the convergence of biology and chemistry and the results may be useful for discovering novel inhibitors and HIF-2α biology studies, and contribute to the development of therapeutic agents.

• KIPS Transactions on Software and Data Engineering
• /
• 제11권12호
• /
• pp.509-516
• /
• 2022

Accurate pose prediction of objects in 3D space is an important visual recognition technique widely used in many applications such as scene understanding in both indoor and outdoor environments, robotic object manipulation, autonomous driving, and augmented reality. Most previous works for object pose estimation have the limitation that they require an exact 3D CAD model for each object. Unlike such previous works, this paper proposes a novel neural network model that can predict the poses of unknown objects based on only their RGB color images without the corresponding 3D CAD models. The proposed model can obtain depth maps required for unknown object pose prediction by using an adaptive depth estimator, AdaBins,. In this paper, we evaluate the usefulness and the performance of the proposed model through experiments using benchmark datasets.