• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2002년도 합동 추계학술대회 논문집 정보 및 제어부문
• /
• pp.342-348
• /
• 2002

This paper presents a computer image processing system, which measures the thickness of the ALC block and inspects the defect on a real-time basis. The Image processing system was established with a CCD camera, an image grabber, and a personal computer without using assembled measurement equipment. The image obtained by the system was analyzed by a devised algorithm, specially designed for the enhanced measurement accuracy. For the realization of proposed algorithm, the pre-processing method that can be applied to overcome uneven lighting environment, and threshold decision method, and subpixel method are developed. from the experimental results, we could find that the required measurement accuracy specification is sufficiently satisfied using our proposed method.

• 소성∙가공
• /
• 제5권4호
• /
• pp.343-352
• /
• 1996

One of the preferred methods that can be used to verify the results of finite element analysis is to measure surface strains of the deformed part for purpose of direct comparison with simulation results. Instead of using the usual manual method the vision-based measurement method is capable of determining surface geometry and strain from the deformed grid pattern automatically with the help of a computer. To obtain strain distribution over an area, the coordinates of such a surface grid are determined from the multiple video images by applying the photogrammetry principle. Methods to improve the overall accuracy of the vision-based strain measurement system are explored and the possible accuracies that can be attained by such a measurement method are discussed. A major emphasis is placed on the initial grid application method its accuracy and ease of subsequent image processing. Finite element analyses of limiting dome height(LDH) test are carried out and the results of them are compared with exsperimen-tal data.

• 디지털융복합연구
• /
• 제3권1호
• /
• pp.149-163
• /
• 2005

Researches for NN(nearest neighbor) query which is often used in LBS system, have been worked. However. Conventional NN query processing techniques are usually meaningless in moving object management system for LBS since their results may be invalidated as soon as the query and data objects move. To solve these problems, in this paper we propose a new nearest neighbor query processing technique, called CTNN, which is possible to meet continuous trajectory nearest neighbor query processing. The proposed technique consists of Approximate CTNN technique which has quick response time, and Exact CTNN technique which makes it possible to search accurately nearest neighbor objects. Experimental results using GSTD datasets shows that the Exact CTNN technique has high accuracy, but has a little low performance for response time. They also shows that the Approximate CTNN technique has low accuracy comparing with the Exact CTNN, but has high response time.

• 한국항공우주학회지
• /
• 제36권6호
• /
• pp.594-599
• /
• 2008

현재 국방과학연구소 수동 시험에서 생성된 영상데이터는 수작업에 의해 분석되어지고 있다. 이러한 방법은 관측자에 따라 정확성과 분석소요시간이 상이할 수 있다 . 본 논문에서는 개선된 영상데이터 처리와 분석을 위해 MATLAB을 기반으로 한 알고리즘을 제안하였다. 이 알고리즘은 왜곡 보정 , 그레이 레벨 변환, 노이즈 제거, 이진화를 하는 영상 전처리 과정, 와류 궤적을 계측하는 영상 분석 과정으로 구성되어 있다 . 수동 시험에서 획득된 영상데이터를 이용하여 테스트 한 결과 제안된 알고리즘은 기존 영상데이터 분석 방법에 비해 정확성과 실행속도가 향상되었다.

• 한국산업융합학회 논문집
• /
• 제22권6호
• /
• pp.627-635
• /
• 2019

In this study, we developed for processing complex features using CAM software that satisfies precision for example practice and related qualification tests suiTable for CNC training purposes. In addition, functions such as location control, speed control, and processing path generation, which are the main functions of CNC machining machines, were constructed using small equipment parts, servo motors, inverters, general purpose PCs, and commercial NC software and researched with the goal of developing low-cost education equipment. In the static accuracy inspection, the degree of machine when measuring the parallelism of the X, Y and Z axes and the vibration of the main shaft did not reach the allowable value. However, we have obtained a finished product that satisfies the CNC machine book sample shape machining, detailed functions of the position control function of the CNC machine tool, linear interpolation function, circular interpolation function, and tool offset function. In the qualification test shape processing, a shape with a degree of 1/100 mm was processed to obtain position accuracy that satisfied the tolerance.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2012년도 춘계학술발표대회
• /
• pp.429-431
• /
• 2012

Vehicle-logo recognition technology is very important in vehicle automatic recognition technique. The intended application is automatic recognition of vehicle type for secure access and traffic monitoring applications, a problem not hitherto considered at such a level of accuracy. Vehicle-logo recognition can improve Vehicle type recognition accuracy. So in this paper, introduces how to vehicle-logo recognition. First introduces the region of the license plate by algorithm and roughly located the region of car emblem based on the relationship of license plate and car emblem. Then located the car emblem with precision by the distance of Hausdorff. On the base, processing the region by morphologic, edge detection, analysis of connectivity and pick up the PCA character by lowing the dimension of the image and unifying the PCA character. At last the logo can be recognized using the algorithm of support vector machine. Experimental results show the effectiveness of the proposed method.

• Journal of Information Processing Systems
• /
• 제15권5호
• /
• pp.1119-1130
• /
• 2019

With the wide spread of Social Network Services (SNS), fake news-which is a way of disguising false information as legitimate media-has become a big social issue. This paper proposes a deep learning architecture for detecting fake news that is written in Korean. Previous works proposed appropriate fake news detection models for English, but Korean has two issues that cannot apply existing models: Korean can be expressed in shorter sentences than English even with the same meaning; therefore, it is difficult to operate a deep neural network because of the feature scarcity for deep learning. Difficulty in semantic analysis due to morpheme ambiguity. We worked to resolve these issues by implementing a system using various convolutional neural network-based deep learning architectures and "Fasttext" which is a word-embedding model learned by syllable unit. After training and testing its implementation, we could achieve meaningful accuracy for classification of the body and context discrepancies, but the accuracy was low for classification of the headline and body discrepancies.

• Journal of Information Processing Systems
• /
• 제15권5호
• /
• pp.1029-1035
• /
• 2019

In recent years, artificial intelligence (AI) services have become one of the most essential parts to extend human capabilities in various fields such as face recognition for security, weather prediction, and so on. Various learning algorithms for existing AI services are utilized, such as classification, regression, and deep learning, to increase accuracy and efficiency for humans. Nonetheless, these services face many challenges such as fake news spread on social media, stock selection, and volatility delay in stock prediction systems and inaccurate movie-based recommendation systems. In this paper, various algorithms are presented to mitigate these issues in different systems and services. Convolutional neural network algorithms are used for detecting fake news in Korean language with a Word-Embedded model. It is based on k-clique and data mining and increased accuracy in personalized recommendation-based services stock selection and volatility delay in stock prediction. Other algorithms like multi-level fusion processing address problems of lack of real-time database.

• 융합신호처리학회논문지
• /
• 제21권4호
• /
• pp.154-161
• /
• 2020

본 논문에서는 인간의 노력을 줄이고 정확성을 높이기 위해 사과의 색을 기반으로 하는 분류 시스템을 제안하였다. 제안된 분류 시스템은 카메라, 모터 및 라즈베리 파이로 구성되어 있고, 미성숙, 성숙, 익은 등으로 총 4가지 종류의 사과를 분류할 수 있다. 시장에서 다양한 종류의 사과를 100개 구입하여 무작위로 선택하여 평가하였다. 정확도는 95%였고 처리 시간은 사과당 약 8초였다. 제안한 시스템은 인력 감축에 유용할 것으로 예상된다.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2021년도 추계학술발표대회
• /
• pp.297-299
• /
• 2021

Text classification is one of the popular tasks in Natural Language Processing (NLP) used to classify text or document applications such as sentiment analysis and email filtering. Nowadays, state-of-the-art (SOTA) Machine Learning (ML) and Deep Learning (DL) algorithms are the core engine used to perform these classification tasks with high accuracy, and they show satisfying results. This paper conducts a benchmarking performance's analysis of multiple SOTA algorithms on the first known labeled Korean voice phishing dataset called KorCCVi. Experimental results reveal performed on a test set of 366 samples reveal which algorithm performs the best considering the training time and metrics such as accuracy and F1 score.