• The Journal of the Acoustical Society of Korea
• /
• v.40 no.5
• /
• pp.466-472
• /
• 2021

Speaker Change Detection (SCD) refers to finding the moment when the main speaker changes from one person to the next in a speech conversation. In speaker change detection, difficulties arise due to overlapping speakers, inaccuracy in the information labeling, and data imbalance. To solve these problems, TIMIT corpus widely used in speech recognition have been concatenated artificially to obtain a sufficient amount of training data, and the detection of changing speaker has performed after identifying overlapping speakers. In this paper, we propose an speaker change detection system that considers the speaker overlapping. We evaluated and verified the performance using various approaches. As a result, a detection system similar to the X-Vector structure was proposed to remove the speaker overlapping region, while the Bi-LSTM method was selected to model the speaker change system. The experimental results show a relative performance improvement of 4.6 % and 13.8 % respectively, compared to the baseline system. Additionally, we determined that a robust speaker change detection system can be built by conducting related studies based on the experimental results, taking into consideration text and speaker information.

• Journal of IKEEE
• /
• v.23 no.2
• /
• pp.586-593
• /
• 2019

In this paper, we introduce the wood pellet electric power generation system, which is an eco - friendly solid fuel processed pure wood, which is one of the largest capacity renewable power fuels in Korea, The Ministry of Commerce, Industry and Energy notified the Ministry of Land, Infrastructure, Transport and Tourism of the Ministry of Land, Transport and Maritime Affairs of the Ministry of Land, Transport and Maritime Affairs. Derived and validated. It is confirmed that the performance of the generator and the voltage control characteristics of excitation system are good even for the change of generator fuel. It can contribute to future reference at the plant that wants to replace fossil fuels with renewable fuels.

• Journal of Aerospace System Engineering
• /
• v.13 no.4
• /
• pp.50-59
• /
• 2019

The design of the main control valve, which is the main component of the flap control system, was based on actual manufacturing experience on the basis of trial and error method. In this paper, a model-based part design method is proosed. The flap control system consists of a main control valve, fail-safe valve, solenoid valve, LVDT and force motor. The main control valve consists mainly of a spool and slot. The important design parameters of the main control valve are the slot width, overlap and clearance. AMESim is linked to the model and it analyzes the flow path of the main control valve. Applying the proposed design procedure, it was confirmed that the required performance was satisfied within the allowable machining error range.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.4
• /
• pp.27-36
• /
• 2022

In this paper, to improve the conditions of the fish processing site, we propose a method to predict the cutting point of fish according to the target weight using AI machine vision. The proposed method performs image-based preprocessing by first photographing the top and front views of the input fish. Then, RANSAC(RANdom SAmple Consensus) is used to extract the fish contour line, and then 3D external information of the fish is obtained using 3D modeling. Next, machine learning is performed on the extracted three-dimensional feature information and measured weight information to generate a neural network model. Subsequently, the fish is cut at the cutting point predicted by the proposed technique, and then the weight of the cut piece is measured. We compared the measured weight with the target weight and evaluated the performance using evaluation methods such as MAE(Mean Absolute Error) and MRE(Mean Relative Error). The obtained results indicate that an average error rate of less than 3% was achieved in comparison to the target weight. The proposed technique is expected to contribute greatly to the development of the fishery industry in the future by being linked to the automation system.

• Journal of IKEEE
• /
• v.26 no.2
• /
• pp.219-225
• /
• 2022

The textile industry in Korea, the dyeing sector is an energy-intensive sector and has low per-unit productivity due to its labor-intensive nature. If the defective rate of dyed fabrics is high, additional costs are incurred due to an increase in production cost due to re-dyeing. Therefore, the goal of the dyeing factory was to minimize the defect rate rather than to save energy. It was difficult to check the dyeing state of the fabric in real time due to the risk of accidents due to burns or pressure when dyeing in a high-temperature and high-pressure environment. In this paper, a complex sensor that can measure the exhaustion rate of dye solution in the dyeing machine using turbidity, pH, and conductivity sensors was proposed, and the experimental method and experimental results were analyzed.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.293-296
• /
• 2021

K-water에서는 다목적댐의 관리를 위해 실시간으로 댐수위, 하천 수위 및 강우량 등을 계측하고 있으며, 계측된 값들은 댐을 효과적으로 운영하는데 필요한 데이터로 활용되고 있다. 특히 댐수위 이상 데이터를 탐지하지 못한 채 그대로 사용할 경우 댐의 방류 시기와 방류량 등을 결정하는 중요한 의사결정을 그르칠 수 있으므로 이를 신속히 탐지하는 것이 매우 중요하다. 현재의 자동화된 이상 데이터 탐지방법 중 하나는 현재 데이터가 최댓값과 최솟값을 초과할 때, 다른 하나는 현재 데이터와 일정 시간 동안의 평균값 간의 차이가 관리자가 정한 특정 값을 벗어났을 때를 기준으로 삼고 있다. 전자는 상한과 하한의 초과 여부만 판단하므로 탐지가 쉬우나 정상범위 내에서 발생한 이상 데이터는 탐지가 불가하다. 후자는 관리자의 경험을 통해 판단 조건을 정하기 때문에 객관성이 결여되는 문제가 있다. 특히 방류와 강우가 복합적으로 댐수위에 영향을 미치는 홍수기에 관리자의 경험에 기초한 이상 데이터 판별은 신뢰성의 문제가 있을 수 있다. 따라서 본 연구에서는 기계학습을 최초로 적용하여 이상 데이터를 탐지하고자 하였다. 댐수위, 누적강우량 및 누적방류량 데이터와 댐수위데이터를 가공하여 생성한 댐수위차, 댐수위차평균, 댐수위평균 등 자질들의 다양한 조합을 만든 후 이를 Random Forest, SVM, AdaptiveBoost 및 다층퍼셉트론(MLP) 등과 같은 여러 가지 기계학습모델 등을 통해 이상 데이터를 판별하는 실험(분류)을 하였다. 실험결과 댐수위, 댐수위차, 댐수위-댐수위평균, 누적강우량, 누적방류량 및 댐수위차평균을 사용하였을 때 MLP에서 가장 우수한 성능을 보였다. 이 연구를 통해서 댐수위 이상 데이터를 기계학습의 분류기능을 통해 효과적으로 탐지할 수 있다는 것과 모델의 성능은 실험에 사용한 자질의 수뿐 아니라 자질의 종류에도 큰 영향을 받는다는 것을 알 수 있었다.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.38 no.7
• /
• pp.935-945
• /
• 2009

This study was performed to find the optimum ratio of ingredients in the Chinese quince jam. The experiment was designed according to the D-optimal design of mixture design, which included 14 experimental points with 4 replicates for three independent variables (Chinese quince paste $45{\sim}60%$, pectin $1.5{\sim}4.5%$, sugar $45.5{\sim}63.5%$). A mathematical analytical tool was employed for the optimization of typical ingredients. The canonical form and trace plot showed the influence of each ingredient in the mixture against final product. By use of F-test, sweetness, pH, L, b, ${\Delta}E$, and firmness were expressed by a linear model, while the spreadmeter value, a, and sensory characteristics (appearance, color, smell, taste, and overall acceptability) were by a quadratic model. The optimum formulations by numerical and graphical method were similar: Chinese quince paste 54.48%, pectin 2.45%, and sugar 53.07%. Optimum ingredient formulation is expected to improve use of Chinese quince and contribute to commercialization of high quality Chinese quince jam.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.51 no.5
• /
• pp.466-476
• /
• 2006

The objective of this study was to compare and analyze the acceptability and consumption attitude for soy foods between Korean and Canadian university students as young consumers. This survey was carried out by questionnaire and the subjects were n=516 in Korea and n=502 in Canada. Opinions for soy foods in terms of general knowledge were that soy foods are healthy (86.5% in Korean and 53.4% in Canadian) or neutral (11.6% in Korean and 42.8% in Canadian), dairy foods can be substituted by soy foods (51.9% in Korean and 41.8% in Canadian), and soy foods are not only for vegetarians and milk allergy Patients but also for ordinary People (94.2% in Korean and 87.6% in Canadian). In main sources of information about soy foods, the rate by commercials on TV, radio or magazine was the highest (58.0%) for Korean students and the rate by family or friend was the highest(35.7%) for Canadian students. In consumption attitude, all of Korean students have purchased soy foods but only 55.4% of Canadian students have purchased soy foods, and soymilk was remarkably recognized and consumed then soy beverage and margarine in order. 76.4% of Korean students and 65.1% of Canadian students think soy foods are general and popular and can purchase easily, otherwise, in terms of price, soy foods were expensively recognized as 'more expensive than dairy foods' was 59.1% (Korean) and 54.7% (Canadian), and 'similar to dairy foods' was 36.8% (Korean) and 39.9% (Canadian). Major reasons for the rare consumption were 'I am not interested in soy foods' in Korean students (27.3%) and 'I prefer dairy foods to soy foods' in Canadian students (51.7%). However, consumption of soy foods in both countries are very positive and it will be increased.

• Applied Chemistry for Engineering
• /
• v.21 no.2
• /
• pp.115-128
• /
• 2010

Recently the use heat sink material grows where the polymer filled with thermal conductive fillers effectively dissipates heat generated from electronic components. Therefore the management of heat is directly related to the lifetime of electronic devices. For the purpose of improving thermal conductivity of composites, fillers with excellent thermaly conductive behavior are commonly used. Polymer composites filled with thermally conductive particles have advantages due to their processibility, cheap price, and durability to the corrosion. This paper aims to review the thermal interface materials and their model equations for predicting the thermal conductivity of polymer composites, and to introduce the commercial thermal conductive fillers and their applications.

• Electronics and Telecommunications Trends
• /
• v.20 no.1 s.91
• /
• pp.139-146
• /
• 2005

폭발적인 정보의 흐름 속에 놓여 있는 지식정보화 시대에서는 필요한 정보를 선택하고 가공하여 새로운지식의 창조, 전달 및 확대 재생산 할 수 있는 능력이 요구되고 있다. 지식정보화 시대가 필요로 하는 창의적 인재 육성을 위한 새로운 교육 형태와 패러다임으로 e-러닝이 유력한 솔루션으로 부각되고 있다.e-러닝은 교수자 중심의 일방향 교육이 아닌 학습자가 시간과 장소에 구애받지 않고 필요한 학습을 하는 양방향 교육 방식이나, 현재까지는 학습 콘텐츠의 단순 반복 학습에 그치고 있어 기존 교실 수업 방식에 비해 학습효율이 떨어지는 점이 이 분야 발전의 장애요소가 되고 있다. 이에 대한 해결책으로 온라인상에서 학습 과제를 여러 명의 학생들이 상호 의존하여 공동으로 해결함으로써 학습 목표를 달성하는 형태의 협력 학습에 대한 연구가 활발히 진행되고 있다. e-러닝 협력학습에는 단순히 학생들을 그룹화하여 함께 학습하도록 하는 것은 교육 효과면 에서 비효율적일 수 있으며, 학생들의 자발적인 참여와 협력을 유도할 수 있는 교수모형을 적용하여야 하며, 개인의 책임감을 근간으로 한 상호 의존적 과제 해결 활동이 이루어져야 한다. 본 고에서는 협력 e-러닝 학습을 위한 연구사례와 기술 동향들에 대해 살펴본다.