• Proceedings of the Korean Statistical Society Conference
• /
• 2000.11a
• /
• pp.173-179
• /
• 2000

Genetic algorithm(GA) is applied for detecting multiple outliers. GA is a heuristic optimization tool solving for near optimal solution. We compare the performance of GA and the other diagnostic measures commonly used for detecting outliers in regression model. The results show that GA seems to have better performance than the others for the detection of multiple outliers.

• Korean Journal of Radiology
• /
• v.22 no.4
• /
• pp.612-623
• /
• 2021

Objective: To evaluate the diagnostic performance of a deep learning algorithm for the automated detection of developmental dysplasia of the hip (DDH) on anteroposterior (AP) radiographs. Materials and Methods: Of 2601 hip AP radiographs, 5076 cropped unilateral hip joint images were used to construct a dataset that was further divided into training (80%), validation (10%), or test sets (10%). Three radiologists were asked to label the hip images as normal or DDH. To investigate the diagnostic performance of the deep learning algorithm, we calculated the receiver operating characteristics (ROC), precision-recall curve (PRC) plots, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) and compared them with the performance of radiologists with different levels of experience. Results: The area under the ROC plot generated by the deep learning algorithm and radiologists was 0.988 and 0.988-0.919, respectively. The area under the PRC plot generated by the deep learning algorithm and radiologists was 0.973 and 0.618-0.958, respectively. The sensitivity, specificity, PPV, and NPV of the proposed deep learning algorithm were 98.0, 98.1, 84.5, and 99.8%, respectively. There was no significant difference in the diagnosis of DDH by the algorithm and the radiologist with experience in pediatric radiology (p = 0.180). However, the proposed model showed higher sensitivity, specificity, and PPV, compared to the radiologist without experience in pediatric radiology (p < 0.001). Conclusion: The proposed deep learning algorithm provided an accurate diagnosis of DDH on hip radiographs, which was comparable to the diagnosis by an experienced radiologist.

• Imaging Science in Dentistry
• /
• v.48 no.4
• /
• pp.245-254
• /
• 2018

Purpose: To assess the effects of object position in the field of view (FOV) and application of a metal artifact reduction (MAR) algorithm on the diagnostic accuracy of cone-beam computed tomography (CBCT) for the detection of vertical root fractures(VRFs). Materials and Methods: Sixty human single-canal premolars received root canal treatment. VRFs were induced in 30 endodontically treated teeth. The teeth were then divided into 4 groups, with 2 groups receiving metal posts and the remaining 2 only having an empty post space. The roots from different groups were mounted in a phantom made of cow rib bone, and CBCT scans were obtained for the 4 different groups. Three observers evaluated the images independently. Results: The highest frequency of correct diagnoses of VRFs was obtained with the object positioned centrally in the FOV, using the MAR algorithm. Peripheral positioning of the object without the MAR algorithm yielded the highest sensitivity for the first observer (66.7%). For the second and third observers, a central position improved sensitivity, with or without the MAR algorithm. In the presence of metal posts, central positioning of the object in the FOV significantly increased the diagnostic sensitivity and accuracy compared to peripheral positioning. Conclusion: Diagnostic accuracy was higher with central positioning than with peripheral positioning, irrespective of whether the MAR algorithm was applied. However, the effect of the MAR algorithm was more significant with central positioning than with peripheral positioning of the object in the FOV. The clinical experience and expertise of the observers may serve as a confounder in this respect.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1997.10a
• /
• pp.247-251
• /
• 1997

FCM(Fuzzy Cognitive Map) is a fuzzy signed directed graph for representing causal reasoning which has fuzziness between causal concepts. Authors have already proposed FCM-based fault diagnostic scheme and verified its usefulness. However, the previously proposed scheme has the problem of lower diagnostic resolution as in the case of other qualitative approaches. In order to improve the diagnostic resolution, a concept of fuzzy number is introduced into the basic FCM-based fault diagnostic algorithm. By incorporation the fuzzy number into fault FCM models, quantitative information such as the transfer gain between the state variables can be effectively utilized for better diagnostic resolution. Furthermore, an enhanced TAM(Temporal Associative Memory) recall procedure and modified and modified pattern matching scheme are also proposed.

• Transactions on Control, Automation and Systems Engineering
• /
• v.1 no.1
• /
• pp.54-61
• /
• 1999

FCM(Fuzzy Cognitive Map) is proposed for representing causal reasoning. Its structure allows systematic causal reasoning through a forward inference. By using the FCM, authors have proposed FCM-based fault diagnostic algorithm. However, it can offer multiple interpretations for a single fault. In process engineering, as experience accumulated, some form of quantitative process knowledge is available. If this information can be integrated into the FCM-based fault diagnosis, the diagnostic resolution can be further improved. The purpose of this paper is to propose an enhanced FCM-based fault diagnostic scheme. Firstly, the membership function of fuzzy set theory is used to integrate quantitative knowledge into the FCM-based diagnostic scheme. Secondly, modified TAM recall procedure is proposed. Considering that the integration of quantitative knowledge into FCM-based diagnosis requires a great deal of engineering efforts, thirdly, an automated procedure for fusing the quantitative knowledge into FCM-based diagnosis is proposed by utilizing self-learning feature of neural network. Finally, the proposed diagnostic scheme has been tested by simulation on the two-tank system.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.9 no.4
• /
• pp.301-308
• /
• 2009

In the modern life, machines are used for various areas in industries as the advance of science and industrial development has proceeded. In many machines, the rotating machines play an important role in many processes. Therefore, the development of fault diagnosis and monitoring system for rotating machines is required. An ubiquitous sensor network (USN) is a combination of the key computer science and engineering area technology including the wireless network, embedded system hardware and software, communication, real-time system, etc. It collects environmental information to realize a variety of functions. In this work, a data reduction algorithm for USN based remote fault diagnostic system which can be easily applied to previously built factories is proposed. To verify the feasibility of the proposed scheme, some simulations and experiments are executed.

• Journal of Biomedical Engineering Research
• /
• v.12 no.1
• /
• pp.63-70
• /
• 1991

Design of Arrhythmia Automatic Diagnostic System Using Decision Table We have developed an arrhythmia automatic diagnostic system using decision table which is based on the criteria of Minnesota code. This system is divided into two Parts. One is wave detection algorithm using significant point extraction method, the other is arrhythmia diag- nostic algorthm. The proposed system allows physicians to diagnose more accurately by pro- viding the objective information about a lot of computer -processed ECG data.

• Smart Structures and Systems
• /
• v.22 no.5
• /
• pp.587-609
• /
• 2018

There is greater significance in identifying the incipient damages in structures at the time of their initiation as timely rectification of these minor incipient cracks can save huge maintenance cost. However, the change in the global dynamic characteristics of a structure due to these subtle damages are insignificant enough to detect using the majority of the current damage diagnostic techniques. Keeping this in view, we propose a hybrid damage diagnostic technique for detection of minor incipient damages in the structures. In the proposed automated hybrid algorithm, the raw dynamic signatures obtained from the structure are decomposed to uni-modal signals and the dynamic signature are reconstructed by identifying and combining only the uni-modal signals altered by the minor incipient damage. We use these reconstructed signals for damage diagnostics using ARMAX model. Numerical simulation studies are carried out to investigate and evaluate the proposed hybrid damage diagnostic algorithm and their capability in identifying minor/incipient damage with noisy measurements. Finally, experimental studies on a beam are also presented to compliment the numerical simulations in order to demonstrate the practical application of the proposed algorithm.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.41 no.2
• /
• pp.53-60
• /
• 2004

This paper is the study for the automatic pulse-power diagnostic system to discriminate the four constitutions with the piezo-sensor module and digital signal processing hardware attached on the patient arm-neck and the statistical decision software instead of the fingers and intelligence of a traditional korean doctor. This system can be used as a important medical equipment because this automatically diagnostic system has shown the excellent performance of the 65∼76% correctness against the 50∼66% correctness which the general korean doctors with knowledge and experiences have shown. Additionally, this paper has discussed the excellent characteristics of the automatic discrimination algorithm of the four constitutions.

• Nuclear Engineering and Technology
• /
• v.36 no.6
• /
• pp.540-548
• /
• 2004

Check valve failure is one of the worst problems in nuclear power plants. Recently, many researches have been based on new technology using accelerometers and ultrasonic and magnetic flux detection have been carried out. Here, we have suggested a method that uses acoustic emission sensors for detecting the failures of check valves through measuring and analyzing backward leakage flow, a system that works without disassembling the check valve. For validating the suggested acoustic emission sensor methodology, we designed a hydraulic test loop with a check valve. We have assumed in this study that check valve failure is caused by disk wear or by the insertion of a foreign object. In addition, we have developed diagnostic algorithms by using a neural network model to identify the type and size of the failure in the check valve. Our results show that the proposed diagnostic algorithm with acoustic emission sensors is a good solution for identifying check valve failure without necessitating any disassembly work.