• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.639-641
• /
• 2023

이미지 기반 폐기물 처리시스템에서 품목별 상이한 수집 난이도로 인해 발생하는 데이터 불균형으로 분류 모델 학습에 어려움이 따른다. 따라서 본 논문에서는 폐기물 분류 모델의 성능 비교를 통해 적합한 이미지 생성 모델을 탐색한다. 데이터의 불균형을 해결할 수 있도록 VAE(Variational Auto-Encoder), GAN(Generative Adversarial Networks) 및 Diffusion Model을 이용하여 이미지를 생성한다. 이후 각각의 생성 방법에 따라 학습데이터와 병합하여 객체 분류를 진행하였다. 정확도는 VAE가 84.41%로 3.3%의 성능 향상을, F1-점수는 Diffusion Model이 91.94%로 6.14%의 성능 향상을 이루었다. 이를 통해, 데이터 수집에서 나타나는 데이터 불균형을 해결하여 실 사용환경에 알맞은 시스템을 구축이 가능함을 확인하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.667-669
• /
• 2023

데이터의 균형은 객체 인식 분야에서 영향을 미치는 요인 중 하나이다. 본 논문에서는 폐기물 데이터 균형을 위해 Chat-GPT와 Diffusion model 기반 데이터 생성 모델을 제안한다. Chat-GPT를 사용하여 폐기물의 속성에 해당하는 단어를 생성하도록 질문하고, 생성된 단어는 인코더를 통해 벡터화시킨다. 이 중 폐기물과 관련 없는 단어를 삭제 후, 남은 단어들을 결합하는 전처리 과정을 거친다. 결합한 벡터는 디코더를 통해 텍스트 데이터로 변환 후, Stable Diffusion model에 입력되어 텍스트와 상응하는 폐기물 데이터를 생성한다. 이 데이터는 AI Hub의 공공 데이터를 활용하며, 객체 인식 모델인 YOLOv5로 학습해 F1-score와 mAP로 평가한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.426-427
• /
• 2023

최근 현대인의 식습관 및 고령화로 인해 당뇨병 환자의 수가 연간 증가하고 있다. 따라서 현재는 아직 당뇨병이 발생하지 않았더라도 미래에 발생할 가능성 예측의 중요성이 커지고 있다. 기존의 당뇨병 발생 여부 진단 연구는 회귀 분석과 같은 단일 모델을 사용하여 수행된다. 그러나 당뇨병에 영향을 미치는 변수들은 복잡하게 얽혀있어 단일 모델만으로는 패턴을 충분히 학습하기 어렵다. 본 논문에서는 데이터에 적합하게 자동으로 다층 스태킹 앙상블 모델을 구성하는 알고리즘을 이용한 다층 스태킹 앙상블 모델을 제안한다. 제안하는 방법은 성능이 높은 모델들을 기준으로 층을 쌓으며 모델을 구성하며 실험 결과 다른 자동 기계학습 라이브러리와 비교해 F1 score 기준으로 최대 12.89%p의 성능 향상을 보였다.

• International journal of advanced smart convergence
• /
• v.6 no.2
• /
• pp.59-64
• /
• 2017

There are three sorts of voices of female vocalists: soprano, mezzo-soprano, and contralto according to the transliteration. Among them, the soprano has the highest vocal range. Since the voice is generated through the human vocal tract based on the voice generation model, it is greatly influenced by the vocal tract. The structure of vocal organs differs from person to person, and the formants characteristic of vocalization differ accordingly. The formant characteristic refers to a characteristic in which a specific frequency band appears distinctly due to resonance occurring in each vocal tract in the vocal process. Formant characteristics include personality that occurs in the throat, jaw, lips, and teeth, as well as phonological properties of phonemes. The first formant is the throat, the second formant is the jaw, the third formant and the fourth formant are caused by the resonance phenomenon in the lips and the teeth. Among them, pronunciation is influenced not only by phonological information but also by jaws, lips and teeth. When the mouth is small or the jaw is stiff when pronouncing, pronunciation becomes unclear. Therefore, the higher the accuracy of the pronunciation characteristics, the more clearly the formant characteristics appear in the grammar spectrum. However, many soprano singers can not open their mouths because their jaws, lips, teeth, and facial muscles are rigid to maintain high tones when singing, which makes the pronunciation unclear and thus the formant characteristics become unclear. In this paper, in order to confirm the accuracy of the pronunciation characteristics of soprano singers, the experimental group was selected as the soprano singers A, B, C, D, E of Korea and analyzed the grammar spectrum and conducted the MOS test for pronunciation recognition. As a result, soprano singer B showed a clear recognition from F1 to F5 and MOS test result showed the highest recognition rate with 4.6 points. Soprano singers A, C, and D appear from F1 to F3, but it was difficult to find formants above 2kHz. Finally, the soprano singer E had difficulty in finding the formant as a whole, and MOS test showed the lowest recognition rate at 2.1 points. Therefore, we confirmed that the soprano singer B, which exhibits the most distinct formant characteristics in the grammar spectrum, has the best pronunciation accuracy.

• Journal of Korean Society for Atmospheric Environment
• /
• v.23 no.3
• /
• pp.344-352
• /
• 2007

Quantifying the methane emission from landfills is important to evaluate measures for reduction of greenhouse gas emissions. To estimate methane emission for the entire landfills from 1990 through 2004 in Korea, Tier 1 and 2 methodologies were used. In addition, five different scenarios were adopted to identify the effect of important variables on methane emission. The trends of methane emission using Tier 1 were similar to the disposed waste amount. Methane emission using Tier 2 increased as the degradation of waste was gradually proceeded. This result indicates that disposed waste amount and methane generation rate are the important variables for the estimation of methane emission by Tier 1 and 2, respectively. As for the different scenarios, methane emission was highest with scenario I that the entire landfills in Korea were regarded as one landfill. Methane emissions by scenario III and IV considering different $DOC_F$ values with the waste type and different MCF values with the height of waste layer, respectively, were underestimated compared to scenario II. This result indicates that the method of scenario I employed to most previous studies may lead to the overestimation of methane emission. Therefore, more careful consideration of the variables should be needed to develop the methodologies of greenhouse gas emission in landfills along with the characteristics of disposed waste in Korea.

• Journal of the Korean Recycled Construction Resources Institute
• /
• v.5 no.3
• /
• pp.305-312
• /
• 2017

This paper presents a laboratory investigation into the effects of fillers using industrial by-product such as coal ash, IGCC slag on properties of hot-mixed asphalt concrete variation with filler content. For comparison, existing mixture with lime and dust have also been considered. Marshall and flow test has been considered for the purpose of mix design as well as evaluation of mixture. Other performance tests such as indirect tensile strength test, tensile strength ratio(moisture susceptibility), dynamic stability have also been carried out variation with filler content. It is observed that the mixes with industrial by-product exhibit conform with quality standard. Therefore, it has been recommended to utilize industrial by-product based on fly ash wherever available, not only reducing the produce cost but also partly solve the industrial by-product utilization and disposal problem.

• Journal of the Korean Solar Energy Society
• /
• v.32 no.3
• /
• pp.146-152
• /
• 2012

To install a wind power generator, the survey on the wind environment resources must be conducted in advance. The survey on the wind environment resources is to collect and analyze data regarding the wind speed and direction on a data logger. The data logger consists of a sensor, signal processing circuit and storage device. According to the analysis of the stored data, the amount of power generation by the types of generators can be predicted and the most optimal generator including safety grade can be selected, and in case of installing a generator in the future, it can be utilized as basic data regarding supporting base and foundation construction method of survey points. Data logger was developed for a small wind power generator that is suitable for the international standard(IEC 61400) by using DSP-F28335 micro controller in this paper. It was developed to measure the wind speed of 1 [m/s]~17 [m/s], the wind direction of 0 [$^{\circ}$]~359 [$^{\circ}$], and temperature of -30 [$^{\circ}C$]~50 [$^{\circ}C$], and the comparative experiment with other companies' data loggers was conducted, and an error was measured to be less than ${\pm}0.1$ [m/s] for wind speed and less than +1 [$^{\circ}$] for wind direction.

• Journal of Photoscience
• /
• v.7 no.3
• /
• pp.87-95
• /
• 2000

As sunlight not always optimized for every terrestrial plant in terms of light quality, quantity and duration, some plants suffer detrimental effects of sunlight exposure under certain conditions. Photoinhibition of photosynthesis is a typical phenomenon representing harmful light effects, commonly observed in many photosynthetic organisms. It is generally accepted that functional, structural loss of photosystem II complex(PSII) is the primary event of photoinhibition. Accumulating data also suggest that singlet oxygen($^1$O$_2$) is the main toxic species directly involved in it. There are two different views on the specific site and mechanism of $^1$O$_2$ production in the photosynthetic membrane. One of them favors the PSII reaction center, where the primary charge pairs recombination occurs as a prerequisite for the generation of $^1$O$_2$, and the other inclines to photosensitized $^1$O$_2$ formation by a substance located outside PSII. This article describes how we, as the advocators of the latter concept, have arrived at the conclusion that $^1$O$_2$ immediately involved in PSII photodamage is largely generated from the Rieske center of the cytochrome b$_{6}$/f complex and diffuses into PSII, attacking the reaction center subunits.s.

• International Journal of Aeronautical and Space Sciences
• /
• v.18 no.1
• /
• pp.48-55
• /
• 2017

Aeroacoustic computation of a fully-developed turbulent pipe flow at $Re_{\tau}=175$ and M = 0.1 is conducted by LES/LPCE hybrid method. The generation and propagation of acoustic waves are computed by solving the linearized perturbed compressible equations (LPCE), with acoustic source DP(x,t)/Dt attained by the incompressible large eddy simulation (LES). The computed acoustic power spectral density is closely compared with the wall shear-stress dipole source of a turbulent channel flow at $Re_{\tau}=175$. A constant decaying rate of the acoustic power spectrum, $f^{-8/5}$ is found to be related to the turbulent bursts of the correlated longitudinal structures such as hairpin vortex and their merged structures (or hairpin packets). The power spectra of the streamwise velocity fluctuations across the turbulent boundary layer indicate that the most intensive noise at ${\omega}^+$ < 0.1 is produced in the buffer layer with fluctuations of the longitudinal structures ($k_zR$ < 1.5).

• Journal of Smart Tourism
• /
• v.1 no.1
• /
• pp.53-64
• /
• 2021

Upselling is often a critical factor in revenue generation for businesses in the tourism and travel industry. Utilizing passenger data from a major international airline company, we develop the PAX (Passenger, Airline, eXternal) model to predict passengers that are most likely to accept an upgrade offer from economy to premium. Formulating the problem as an extremely unbalanced, cost-sensitive, supervised binary classification, we predict if a customer will take an upgrade offer. We use a feature vector created from the historical data of 3 million passenger records from 2017 to 2019, in which passengers received approximately 635,000 upgrade offers worth more than $422,000,000 U.S. dollars. The model has an F1-score of 0.75, outperforming the airline's current rule-based approach. Findings have several practical applications, including identifying promising customers for upselling and minimizing the number of indiscriminate emails sent to customers. Accurately identifying the few customers who will react positively to upgrade offers is of paramount importance given the airline 'industry's razor-thin margins. Research results have significant real-world impacts because there is the potential to improve targeted upselling to customers in the airline and related industries.