• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.01a
• /
• pp.9-12
• /
• 2023

최근 영상 콘텐츠를 통해 영상물의 메시지뿐 아니라 메시지의 형식을 통해 전달된 감정이 시청하는 사람의 심리 상태에 영향을 주고 있다. 이에 따라, 영상 콘텐츠의 감정을 분류하는 연구가 활발히 진행되고 있고 본 논문에서는 대중적인 영상 스트리밍 플랫폼 중 하나인 유튜브 영상을 7가지의 감정 카테고리로 분류하는 여러 개의 영상 데이터 중 각 영상 데이터에서 오디오와 이미지 데이터를 각각 추출하여 학습에 이용하는 멀티 모달 방식 기반의 영상 감정 분류 모델을 제안한다. 사전 학습된 VGG(Visual Geometry Group)모델과 ViT(Vision Transformer) 모델을 오디오 분류 모델과 이미지 분류 모델에 이용하여 학습하고 본 논문에서 제안하는 병합 방법을 이용하여 병합 후 비교하였다. 본 논문에서는 기존 영상 데이터 감정 분류 방식과 다르게 영상 속에서 화자를 인식하지 않고 감정을 분류하여 최고 48%의 정확도를 얻었다.

• Journal of Astronomy and Space Sciences
• /
• v.40 no.4
• /
• pp.259-266
• /
• 2023

The Standard Model (SM) does not provide an information for 26% of dark matter of the universe. In the dark sector, dark matter is supposed to be linked with the hypothetical particles called dark photons that have similar role to photons in electromagnetic interaction in the SM. Besides astronomical observation, there are studies to find dark matter candidates using accelerators. In this paper, we searched for dark photons using future electron-positron colliders, including Circular Electron Positron Collider (CEPC)/CEPC, Future Circular Collider (FCC-ee)/Innovative Detector for Electron-positron Accelerator (IDEA), and International Linear Collider (ILC)/International Large Detector (ILD). Using the parameterized response of the detector simulation of Delphes, we studied the sensitivity of a double dark photon mode at each accelerator/detector. The signal mode is double dark photon decay channel, e⁺e^- → A'A', where A' (dark photon with spin 1) decaying into a muon pair. We used MadGraph5 to generate Monte Carlo (MC) events by means of a Simplified Model. We found the dark photon mass at which the cross-sections were the highest for each accelerator to obtain the maximum number of events. In this paper we show the expected number of dark photon signal events and the detector efficiency of each accelerator. The results of this study can facilitate in the dark photon search by future electron-positron accelerators.

• Journal of Information Management
• /
• v.43 no.2
• /
• pp.169-197
• /
• 2012

Patents are an up-to-date and reliable source of technological knowledge, and thus patent analysis has been considered to be a necessary step for identifying evolving technological trends and planning technology strategies. Although there exist many research papers and technical reports concerning patent analysis, few empirical studies on planning technology strategies for uncertain futures from a national or company perspective have been rarely conducted. Therefore, this paper aims presenting a procedure and its practical case of planning future technology strategies by incorporating patent analysis and scenario planning. Using patents related to polymer electrolyte membrane fuel cells, this paper developed technology strategies corresponding to future scenarios. We expect that the proposed method and case study can assist knowledge services of experts in the long-term technology strategy planning process.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2009.11a
• /
• pp.99-99
• /
• 2009

Two different micro-flows during the evaporation of ink droplets were achieved by engineering both surface tension gradient and compositional gradient across the ink droplet: (1) Coffee-ring generating flow resulting from the outward flow inside the ink droplet & (2) Marangoni flow leading to the circulation flow inside the ink droplet. The surface tension gradient and the compositional gradient in the ink droplets were tailored by mixing two different solvents with difference surface tension and boiling point. In order to create the coffee-ring generating flow (outward flow), a single-solvent system using N,N-dimethylformamide with nano-sized spherical alumina particles was formulated, Marangoni flow (circulation flow) was created in the ink droplets by combining N,N-dimethylformamide and fotmamide with the spherical alumina powders as a co-solvent ink system. We have investigated the effect of these two different flows on the formation of ceramic films by inkjet printing method, The packing density of the ceramic films printed with two different ink systems (single- and co-solvent systems) and their surface roughness were characterized. The dielectric properties of these inkjet-printed ceramic films such as dielectric constant and dissipation factor were also studied in order to evaluate the feasibility of their application to the electronic ceramic package substrate.

• Mass Spectrometry Letters
• /
• v.3 no.1
• /
• pp.21-24
• /
• 2012

$5{\alpha}$-Dihydrotestosterone (DHT) is the primary active metabolite of testosterone, catalyzed by $5{\alpha}$-reductase ($5{\alpha}R$) in the skin, prostate, and liver. In this study, the $5{\alpha}R$ activity in rat liver S9 fraction in the presence of a NADPH-generating system was evaluated and compared by gas chromatography-mass spectrometry (GC-MS)-based in vitro assays. Testosterone and a $5{\alpha}R$ inhibitor, finasteride, were added to the S9 fractions and incubated at $37^{\circ}C$ for 1 h. Both testosterone and DHT were quantitatively measured and compared with two different GC-MS-based steroid profiling techniques. DHT was not detected by conventional GC-MS analysis in the absence of finasteride when the concentration of testosterone in the S9 fraction was less than $0.2{\mu}M$, whereas the isotope-dilution GC-MS (GC-IDMS) system was able to evaluate the $5{\alpha}R$ activity. Because the S9 fraction contains more reactive enzymes and is easier to collect from tissues compared with a microsomal solution, the combination of the S9 fraction and GC-IDMS technique may be a promising assay for evaluating the $5{\alpha}R$ activity in large-scale clinical studies.

• Journal of Broadcast Engineering
• /
• v.21 no.4
• /
• pp.506-514
• /
• 2016

In this study, HEVC over DVB-T2 systems with a bandwidth of 6 MHz is considered, particularly for the terrestrial 4K-UHDTV broadcasting service in the Republic of Korea. The threshold of visibility carrier-to-noise power ratio (ToV C/N) and the receiver minimum required input level (sensitivity) for satisfying the subjective picture failure (SPF) condition are measured in the laboratory. It is observed, for transmitting 26.37 Mbps data stream correctly, that ToV C/N is 18.8 dB on average, and the receiver sensitivity is varied from minimum -84.2 dBm to maximum -80.0 dBm. Based on the results, the receiver noise floor is calculated by -100 dBm on average.

• 한국균학회소식:학술대회논문집
• /
• 2014.10a
• /
• pp.19-19
• /
• 2014

Pleurotus ostreatus, P. eryngii, and Flammulina velutipes are major edible mushrooms that account for over 89% of total mushroom production in Korea. Recently, Agrocybe cylindracea, Hypsizygus marmoreus, and Hericium erinaceu are increasingly being cultivated in mushroom farms. In Korea, the production of edible mushrooms was estimated to be 614,224 ton in 2013. Generally, about 5 kg of mushroom substrate is needed to produce 1 kg of mushroom, and consequently about 25 million tons of spent mushroom substrate (SMS) is produced each year in Korea. Because this massive amount of SMC is unsuitable for reuse in mushroom production, it is either used as garden fertilizer or deposited in landfills, which pollutes the environment. It is reasonably assumed that SMS includes different secondary metabolites and extracellular enzymes produced from mycelia on substrate. Three major groups of enzymes such as cellulases, xylanases, and lignin degrading enzymes are involved in breaking down mushroom substrates. Cellulase and xylanase have been used as the industrial enzymes involving the saccharification of biomass to produce biofuel. In addition, lignin degrading enzymes such as laccases have been used to decolorize the industrial synthetic dyes and remove environmental pollutions such as phenolic compounds. Basidiomycetes produce a large number of biologically active compounds that show antibacterial, antifungal, antiviral, cytotoxic or hallucinogenic activities. However, most previous researches have focused on therapeutics and less on the control of plant diseases. SMS can be considered as an easily available source of active compounds to protect plants from fungal and bacterial infections, helping alleviate the waste disposal problem in the mushroom industry and creating an environmentally friendly method to reduce plant pathogens. We describe extraction of lignocellulytic enzymes and antimicrobial substance from SMSs of different edible mushrooms and their potential applications.

• KEPCO Journal on Electric Power and Energy
• /
• v.2 no.1
• /
• pp.109-113
• /
• 2016

Techno-economic evaluation of Non-Capture $CO_2$ Utilization (NCCU) technology for the production of high-value-added products using greenhouse gas ($CO_2$) was performed. The general scheme of NCCU process is composed of $CO_2$ carbonation and brine electrolysis process. Through a carbonation reaction with sodium hydroxide that is generated from brine electrolysis and $CO_2$ of the flue gas, it is possible to get high-value-added products such as sodium bicarbonate, sodium hydroxide, hydrogen & chloride and also to reduce the $CO_2$ emission simultaneously. For the techno-economic study on NCCU technology, continuous operation of bench-scale facility which could treat $2kgCO_2/day$ was performed. and based on the key performance data evaluated, the economic evaluation analysis targeted on the commercial chemical plant, which could treat 6 tons $CO_2$ per day, was performed using the net present value (NPV) metrics. The results showed that the net profit obtained during the whole plant operation was about 7,890 mKRW (million Korean Won) on NPV metrics and annual $CO_2$ reduction was estimated as about $2,000tCO_2$. Also it was found that the energy consumption of brine electrolysis is one of the key factors which affect the plant operation cost (ex. electricity consumption) and the net profit of the plant. Based on these results, it could be deduced that NCCU technology of this study could be one of the cost-effective $CO_2$ utilization technology options.

• Industrial Engineering and Management Systems
• /
• v.15 no.3
• /
• pp.197-205
• /
• 2016

The public services closely related to the daily lives of the Japanese people, such as firefighting, police or primary school education, are largely financed by the local governments. As the population as a whole in Japan declines, the population in local regions are forecasted to experience particularly rapid decline in the future, and it is inevitable to reduce the cost of public services provided by the local governments to keep their financial basis sustainable. In order to provide public services to the people properly and fairly, the local governments own and utilize their public facilities, such as fire stations, police stations or primary schools. On the other hand, we have to secure the accessibility, which is the condition of accessing a facility easily in a whole local city including the high population density area and low population density area. In this paper, we propose a method of determining the number of future facilities and its facility locations in which we maintain the present accessibility. In our proposed method, we determine them comparing the accessibility measurement calculated by facility location model using the present and future population. We adopted k-centdian model as the facility location model, which can secure the accessibility in a whole local city determining the weights of both areas. We applied our proposed method to fire station in Iwaki city, Japan. The results suggested that 7 facilities would be reduced in 2064, after 50 years from 2014. Additionally, we confirmed that the future facility location had secured accessibility in both high and low population density area.

• Journal of the Society of Disaster Information
• /
• v.18 no.2
• /
• pp.364-373
• /
• 2022

Purpose: The purpose of this paper is to develope smoke detection using AI model for detecting the initial fire in underground utility tunnels using CCTV Method: To improve detection performance of smoke which is high irregular, a deep learning model for fire detection was trained to optimize smoke detection. Also, several approaches such as dataset cleansing and gradient exploding release were applied to enhance model, and compared with results of those. Result: Results show the proposed approaches can improve the model performance, and the final model has good prediction capability according to several indexes such as mAP. However, the final model has low false negative but high false positive capacities. Conclusion: The present model can apply to smoke detection in underground utility tunnel, fixing the defect by linking between the model and the utility tunnel control system.