• Journal of Intelligence and Information Systems
• /
• v.28 no.4
• /
• pp.329-346
• /
• 2022

Recently, as the market for short form videos (Instagram, TikTok, YouTube) on social media has gradually increased, research using them is actively being conducted in the artificial intelligence field. A representative research field is Video to Shop, which detects fashion products in videos and searches for product images. In such a video-based artificial intelligence model, product features are extracted using convolution operations. However, due to the limitation of computational resources, extracting features using all the frames in the video is practically impossible. For this reason, existing studies have improved the model's performance by sampling only a part of the entire frame or developing a sampling method using the subject's characteristics. In the existing Video to Shop study, when sampling frames, some frames are randomly sampled or sampled at even intervals. However, this sampling method degrades the performance of the fashion product search model while sampling noise frames where the product does not exist. Therefore, this paper proposes a sampling method MF (Missing Fashion items on frame) sampler that removes noise frames and improves the performance of the search model. MF sampler has improved the problem of resource limitations by developing a keyframe mechanism. In addition, the performance of the search model is improved through noise frame removal using the noise detection model. As a result of the experiment, it was confirmed that the proposed method improves the model's performance and helps the model training to be effective.

• Ecology and Resilient Infrastructure
• /
• v.1 no.2
• /
• pp.102-108
• /
• 2014

Studies of non-point pollutant treatment facilities have widely been conducted for a decade, but natural non-point pollutant treatment facilities implemented on roads have not been carried out for the removal efficiency of non-point pollution sources. This study analyzed the removal efficiency of non-point pollution sources from constructed wetlands using monitoring and event mean concentration method. As a result of this study, removal efficiency of general non-point pollution sources as TSS, COD, BOD is relatively good, but removal efficiency of TN, TP, Cr, Zn, Pb is very small or nothing.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.25-25
• /
• 2020

최근 기후변화와 함께 한국은 매년 가뭄이 지속적으로 발생하고 있으며 피해지역도 다양하게 나타나고 있다. 이에, 가뭄에 능동적으로 대처하고, 지하수자원을 확보할 수 있는 인공함양 및 취수시설의 실증시설을 구축하여 상시 가뭄지역을 대상으로 적용하고자 한다. 최종적으로 연구지역의 소규모 평탄지 하천 중·상류지역 충적층에 인공함양을 통해 가뭄대책 마련을 위한 인공 주입시설을 설치하고, 대수층 발달이 양호한 하류지역에 수평집수정을 이용한 취수시설을 설치하여 취수-공급의 지속적인 순환방식의 통합적인 인공주입·취수시스템을 구축하고자 한다. 연구지역은 가뭄 및 상수도 미 보급지역으로 인해 제한급수가 이루어지고 있는 충청남도 홍성군 운곡리 지역을 대상으로 하였으며, 연구지역 주 대수층인 충적층은 시추조사 결과 지표하 0.5~12.3m에 분포하며, 실트질 모래 및 실트질 자갈로 구성되어있다. 대수층의 수리성은 양수 및 회복시험결과 3.79×10-4~7.71×10^-3(평균 3.13×10^-3)cm/sec로 인공함양을 위한 지반 특성으로 양호한 수리특성을 보인다. 본 연구에서는 연구지역의 지층분포 및 수리지질특성을 고려하여 현장시험 대수층에 인공함양 방법 중 Ditch 및 Ditch+injection Well을 설치하고 공경, 여재등 여러인자에 따른 인공 주입에 의한 함양특성과 수위변동 모니터링을 분석하였으며, 분석결과를 토대로 향후 최적의 취수시설을 구축하여 국내 상시 가뭄지역의 대체수자원확보와 용수공급 방안으로 활용할 수 있을 것으로 판단된다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.12
• /
• pp.1872-1879
• /
• 2022

Recently, with the development of artificial intelligence technology, research to use artificial intelligence to detect hacking attacks is being actively conducted. However, the fact that security data is a representative imbalanced data is recognized as a major obstacle in composing the learning data, which is the key to the development of artificial intelligence models. Therefore, in this paper, we propose a W-VAE oversampling technique that applies VAE, a deep learning generation model, to data extraction for oversampling, and sets the number of oversampling for each class through weight calculation using K-NN for sampling. In this paper, a total of five oversampling techniques such as ROS, SMOTE, and ADASYN were applied through NSL-KDD, an open network security dataset. The oversampling method proposed in this paper proved to be the most effective sampling method compared to the existing oversampling method through the F1-Score evaluation index.

• Bulletin of the Korean Space Science Society
• /
• 1994.09a
• /
• pp.11-11
• /
• 1994

• Proceedings of the Korean Magnestics Society Conference
• /
• 1994.03a
• /
• pp.28-29
• /
• 1994

• Journal of Wetlands Research
• /
• v.13 no.3
• /
• pp.481-488
• /
• 2011

Various development activities have lead to the destruction of the ecosystem such as natural wetlands. In order to protect these natural wetlands, the Ministry of Environment (MOE) in Korea enacted the Wetland Conservation Act in 1999 and designated protected areas for wetland conservation. The MOE adapted the use of Best Management Practices (BMP) such as retention ponds and constructed wetlands to treat the polluted water before entering the water system. One of these projects was a free-water surface flow (FWS) constructed wetland built as a secondary treatment unit for piggery wastewater effluent coming from a livestock wastewater treatment facility. Water quality monitoring for the constructed wetland was conducted during rainfall events. The results showed that the average removal efficiencies of TSS, BOD, TN, TP were 86, 60, 45, 70%, respectively. It was observed that the removal efficiency of particulate matter and phosphorus was high compared to nitrogen. Therefore, a longer hydraulic retention time was needed in order to improve the treatment efficiency of nitrogen. The results of this study can contribute to the wetland design, operation and maintenance of constructed wetlands.

• Proceedings of the Korea Air Pollution Research Association Conference
• /
• 2003.05b
• /
• pp.413-414
• /
• 2003

최근의 대기오염과 기후변화는 특정 지역에만 국한되지 않고 전지구 규모로 발생하고 있다. 따라서 인공위성에 탑재된 원격 센서들을 이용한 대기환경 모니터링이 주목을 받고 있다. 일산화탄소 (CO)는 OH 농도와 직접적인 관련이 있는 대류권 화학에서 매우 중요한 미량기체이며, 대기 중 lifetime이 약 2개월이므로 산불이나 대규모 공업단지에서 생성된 CO를 포함한 오염물질들의 추적자로 사용될 수 있다. 이러한 취지에서 MOPITT (Measurement of Pollution in The Troposphere) 기기가 개발되어, 1999년에 지구관측위성인 Terra에 탑재되어 CO 및 CH$_4$ 모니터링을 수행하고 있다. (중략)

• Journal of the Korean Magnetics Society
• /
• v.2 no.3
• /
• pp.251-256
• /
• 1992

Artificial superlattices of Co/Pd were prepared by RF magnetron sputtering Multilayered structure and compositional modulation were analyzed with a side angle x-ray diffractometer. It has been found that expansion of Co lattice occured in this artificial superlattice due to the lattice mismatch between Co and Pd. Perpendicular magnetic anisotropy could be observed when the Co layer thickness became less than 8${\AA}$ and maximum coercivity of 2350 Oe could be obtained in [Co(2.5 ${\AA}$)/Pd(9.3 ${\AA})]_{50}$/Pd$(200\;{\AA})$ with a perfect squareness of magnetic hysteresis loop. Characteristic of perpendicular magnetic anisotropy in Co/Pd superlattices could be related to the expansion of Co lattice caused by Pd layer and it turned out that as the thickness of Pd layer increased, perpendicular magnetic anisotropy increased. The interface anisotropy energy and volume anisotropy energy were calculated to be 0.29 ergs/$cm^2$ and -$6.9{\times}10^6$ ergs/$cm^3$ respectively, which are consistent with the values reported elsewhere.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.23 no.1
• /
• pp.116-124
• /
• 2009

This paper proposes a new Artificial-Intelligent(AI) Power Quality(PQ) diagnosis algorithm using Discrete Wavelet Transform(DWT), Fast Fourier Transform(FFT), Root-Mean-Square(RMS) value. The developed algorithm is able to detect and classify the PQ problems such as the transient, the voltage sag, the voltage swell, the voltage interruption and the total harmonics distortion. The 15.36[kHz] sampling frequency is used to measure the voltages in a power system. The measured signals are used for DWT, FFT, RMS calculation. For AI diagnosis of the PQ problems, a simple multi-layered Artificial Neural Network(ANN) with the back-propagation algorithm is adopted, programmed in C++ and tested in PSIM simulation studies. Finally, the algorithm, which is installed in MP PQ+256 with TI DSP320C6713, is proved to diagnose the PQ problems efficiently.