• Journal of the Korean Society of Radiology
• /
• v.16 no.5
• /
• pp.595-602
• /
• 2022

This study tried to suggest the most efficient learning rate for accurate and efficient automatic diagnosis of medical images for chest X-ray pneumonia images using deep learning. After setting the learning rates to 0.1, 0.01, 0.001, and 0.0001 in the Inception V3 deep learning model, respectively, deep learning modeling was performed three times. And the average accuracy and loss function value of verification modeling, and the metric of test modeling were set as performance evaluation indicators, and the performance was compared and evaluated with the average value of three times of the results obtained as a result of performing deep learning modeling. As a result of performance evaluation for deep learning verification modeling performance evaluation and test modeling metric, modeling with a learning rate of 0.001 showed the highest accuracy and excellent performance. For this reason, in this paper, it is recommended to apply a learning rate of 0.001 when classifying the presence or absence of pneumonia on chest X-ray images using a deep learning model. In addition, it was judged that when deep learning modeling through the application of the learning rate presented in this paper could play an auxiliary role in the classification of the presence or absence of pneumonia on chest X-ray images. In the future, if the study of classification for diagnosis and classification of pneumonia using deep learning continues, the contents of this thesis research can be used as basic data, and furthermore, it is expected that it will be helpful in selecting an efficient learning rate in classifying medical images using artificial intelligence.

• Proceedings of the KIEE Conference
• /
• 1999.07c
• /
• pp.1493-1495
• /
• 1999

본 논문에서는 변압기 유중 가스 분석 자료와 고장에 관련된 특징을 학습시킨 신경 회로망을 이용하여 전력용 유입 변압기의 새로운 고장 진단 방법을 제안하였다. 본 논문에서 제안한 신경 회로망을 이용한 고장 진단 방법(유중 가스 분석 방법)은 입력으로 가스 구성비 분석(IEC 기준) 및 주요 가스 분석(한국 전력 공사 기준) 자료를 채택하였다. 또한, 출력으로 전력용 유입 변압기의 고장 유무 및 고장 종류의 특징을 신경 회로망으로 추출하였다. 따라서 입력된 유중 가스 분석 결과에 따라 전력용 유입 변압기의 진단 결과(고장 유무 인식 및 해석)가 제시되도록 구성하였다. 제안된 신경회로망을 이용한 변압기 고장 진단 방법은 한국 전력 공사의 변압기 유중 가스 기록으로 효용성을 입증하였다. 따라서 유중 가스 분석만으로 현실성 있는 변압기 진단 및 상태 추정이 가능하게 되었고, 이것의 적용으로 적절한 유지 및 보수 대책 방안을 제시할 수 있게 되었다.

• Proceedings of the KIEE Conference
• /
• 2004.07d
• /
• pp.2245-2247
• /
• 2004

본 논문에서는 미지의 유계를 가진 불확실성을 포함한 비선형 시스템에 대한 고장 진단 설계를 제안한다. 제안된 고장 진단 필터는 비선형 관측기 설계 기술에 기초하여 설계되며, 신경망을 이용하여 고장 성분과 불확실성 성분을 추정하고 추정된 불확실성의 상한값을 고장 진단에 이용한다. 제안된 근사기는 불확실성과 고장 함수를 추정함으로써 고장 검출뿐만 아니라 고장 진단을 확인할 수 있도록 설계된다. 모의실험을 통해서 제안된 고장 진단 설계의 성능을 검증하였다.

• Proceedings of the KIEE Conference
• /
• 1996.07b
• /
• pp.1365-1367
• /
• 1996

본 연구에서는 확정진단시의 무릎 CT 화상을 대상으로해서 화상인식의 제1단계로써, 최근 여러분야에서 많은연구가 행해지고 있는 3층역전파신경회로망(BPN)에 의한 무릅덮개뼈 탈구중의 자동진단 가능성에 관해서 검토를 행했다. 실험결과로부터 신경회로망에 의한 무릅덥개뼈 탈구화상의자동진단은 충분하다고 할 수는 없어도 가능성이 있음올 알게 되었다. 다만, 본 실험에서사용된 패턴수가 적어, 충분한 학습이 이루어지지 않았을 가능성이 있으며, 또한 test된 화상수도 충분치 못하였다. 데이터의 증가에 수반해서 인식률이 충분히 한 향상될 것으로사료된다. 신경회로망은 원리척으로 패턴변환의 한 종류로써, 현상태의 기술수준을 고려할때 과도의 기대는 금물이지만, 패턴인식, 화상처리 등 종래의 계산기가 능숙하게 대처하지 못했던 분야에 대해서 큰 기대를 부여하고 있다. 특히 의공학연구에 있어서 BPN의 응용범위를 사고한다면, 확정진단시에 있어 의사가 보다 확실한 진단을 할 수 있도록 진단지원에 휴익한 도움을 줄 수 있을 것으로 사료된다.

• The Journal of Bigdata
• /
• v.2 no.2
• /
• pp.75-86
• /
• 2017

In this study, I propose a method to automate the method to diagnose the quality of big data. The reason for automating the quality diagnosis of Big Data is that as the Fourth Industrial Revolution becomes a issue, there is a growing demand for more volumes of data to be generated and utilized. Data is growing rapidly. However, if it takes a lot of time to diagnose the quality of the data, it can take a long time to utilize the data or the quality of the data may be lowered. If you make decisions or predictions from these low-quality data, then the results will also give you the wrong direction. To solve this problem, I have developed a model that can automate diagnosis for improving the quality of Big Data using machine learning which can quickly diagnose and improve the data. Machine learning is used to automate domain classification tasks to prevent errors that may occur during domain classification and reduce work time. Based on the results of the research, I can contribute to the improvement of data quality to utilize big data by continuing research on the importance of data conversion, learning methods for unlearned data, and development of classification models for each domain.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.11
• /
• pp.1519-1524
• /
• 2021

In this paper, we propose a method that uses fuzzy decision tree based health pre-diagnosis system of oriental medicine. The proposed fuzzy decision tree based health pre-diagnosis system uses the data from the past which has been pre-trained to get the boundary values based on entropy then, when the user inputs the symptoms, the top 5 diseases that causes those symptoms are extracted. With the extracted top 5 diseases, the system provides information on those diseases with the cause and how to treat them with folk remedies. The database of the diseases and their symptoms is established with the information based on the various books that the oriental doctor recommended then reviewed by the oriental doctor for confirmation. By utilizing the data from the past to train the symptoms of the diseases, the proposed oriental medicine-based health pre-diagnosis system method could provide more accurate diagnosis results faster.

• Journal of Korea Society of Industrial Information Systems
• /
• v.27 no.6
• /
• pp.95-103
• /
• 2022

Quick and accurate understanding of the situation in a severe accident is essential for conducting the appropriate accident management and response using the accident diagnosis information. This study employed deep learning technology to diagnose severe accidents through the major safety parameters transferred from a nuclear power plant (NPP) to AtomCARE. After selecting the major accident scenarios to consider, a learning database was established for particular scenarios affiliated with major scenarios by performing a large number of severe accident analyses using MAAP5 code. The severe accident diagnosis technology, which classifies detailed accident scenarios using the major safety parameters from NPPs, was developed by training it with the established database . Verification and validation were conducted by blind test and principal component analysis. The technology developed in this study is expected to be extended and applied to all severe accident scenarios and be utilized as a base technology for quick and accurate severe accident diagnosis.

• Asia-pacific Journal of Multimedia Services Convergent with Art, Humanities, and Sociology
• /
• v.8 no.3
• /
• pp.571-578
• /
• 2018

The deteriorated pipline often causes the unexpected leakage and crack. Negligence and late maintenance leads the enormous damage for gas and water resource. This paper proposes early fault detection and diagnosis algorithm for pipeline using acoustic emission (AE) signals. Early fault detection method for pipeline compares the frequency amplitude of the spectrum to that of the spectrum in normal condition. Larger amplitude of the spectrum indicates abnormal condition. Early fault diagnosis algorithm uses support vector machines (SVM), which is trained for normal and abnormal conditions to diagnose the measured AE signal from the target pipeline. In the experiment, a pipeline testbed is constructed similarly to real industrial pipeline. Normal, 5mm cracked, 10mm holed pipelines are installed and tested in this study. The proposed fault detection and diagnosis technique is validated as an efficient approach to detect early faulty condition of pipeline.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.11
• /
• pp.153-161
• /
• 2013

Recently, there has been a tremendous increase in the availability of educational materials for foreign language learning. As part of this trend, there has been an increase in the amount of electronically mediated materials available. However, conventional educational contents developed using computer technology has provided typically one-way information, which is not the most helpful thing for users. Providing the user's convenience requires additional off-line analysis for diagnosing an individual user's learning. To improve the user's comprehension of texts written in a foreign language, we propose an intelligent learning tool based on the analysis of the user's eyeball movements, which is able to diagnose and improve foreign language reading ability by providing necessary supplementary aid just when it is needed. To determine the user's learning state, we correlate their eye movements with findings from research in cognitive psychology and neurophysiology. Based on this, the learning tool can distinguish whether users know or do not know words when they are reading foreign language sentences. If the learning tool judges a word to be unknown, it immediately provides the student with the meaning of the word by extracting it from an on-line dictionary. The proposed model provides a tool which empowers independent learning and makes access to the meanings of unknown words automatic. In this way, it can enhance a user's reading achievement as well as satisfaction with text comprehension in a foreign language.

• The Journal of the Korea Contents Association
• /
• v.7 no.6
• /
• pp.161-168
• /
• 2007

In this paper, we proposed the mathematics learning evaluation system between teachers and students using the web. The proposed web-based evaluation system lets learners make up their lesson in a self-oriented and effective way, by letting instructors diagnose learners level of understanding learned contents and letting learners take part in evaluation as well. The system also lets instructors easily make out items for evaluation by using hangul(word processor) and present them on the web. With the help of this web-based mathematics learning site and mathematics learning evaluation system, learners can perform self-oriented loaming and approach various kinds of problems. In addition, students can check with answers and have feedback on the spot. As a result, students can solve lack of understanding on the learned contents.