• Journal of Periodontal and Implant Science
• /
• v.48 no.2
• /
• pp.114-123
• /
• 2018

Purpose: The aim of the current study was to develop a computer-assisted detection system based on a deep convolutional neural network (CNN) algorithm and to evaluate the potential usefulness and accuracy of this system for the diagnosis and prediction of periodontally compromised teeth (PCT). Methods: Combining pretrained deep CNN architecture and a self-trained network, periapical radiographic images were used to determine the optimal CNN algorithm and weights. The diagnostic and predictive accuracy, sensitivity, specificity, positive predictive value, negative predictive value, receiver operating characteristic (ROC) curve, area under the ROC curve, confusion matrix, and 95% confidence intervals (CIs) were calculated using our deep CNN algorithm, based on a Keras framework in Python. Results: The periapical radiographic dataset was split into training (n=1,044), validation (n=348), and test (n=348) datasets. With the deep learning algorithm, the diagnostic accuracy for PCT was 81.0% for premolars and 76.7% for molars. Using 64 premolars and 64 molars that were clinically diagnosed as severe PCT, the accuracy of predicting extraction was 82.8% (95% CI, 70.1%-91.2%) for premolars and 73.4% (95% CI, 59.9%-84.0%) for molars. Conclusions: We demonstrated that the deep CNN algorithm was useful for assessing the diagnosis and predictability of PCT. Therefore, with further optimization of the PCT dataset and improvements in the algorithm, a computer-aided detection system can be expected to become an effective and efficient method of diagnosing and predicting PCT.

• Asian Pacific Journal of Cancer Prevention
• /
• v.16 no.1
• /
• pp.345-349
• /
• 2015

Background: To determine the cut-off values of the preoperative risk of malignancy index (RMI) used in differentiating benign or malignant adnexal masses and to determine their significance in differential diagnosis by comparison of different systems. Materials and Methods: 191 operated women were assessed retrospectively. RMI of 1, 2, 3 and 4; cut-off values for an effective benign or malignant differentiation together with sensitivity, specificity, negative and positive predictive values were calculated. Results: Cut-off value for RMI 1 was found to be 250; there was significant (p<0.001) compatibility at this level with sensitivity of 60%, positive predictive value (PPV) of 75%, specificity of 93%, negative predictive value (NPV) of 88% and an overall compliance rate of 85%. When RMI 2 and 3 was obtained with a cut-off value of 200, there was significant (p<0.001) compatibility at this level for RMI 2 with sensitivity of 67%, PPV of 67%, specificity of 89%, NPV of 89%, histopathologic correlation of 84% while RMI 3 had significant (p<0.001) compatibility at the same level with sensitivity of 63%, PPV of 69%, specificity of 91%, NPV of 88% and a histopathologic correlation of 84%. Significant (p<0.001) compatibility for RMI 4 with a sensitivity of 67%, PPV of 73%, specificity of 92%, NPV of 89% and a histopathologic correlation of 86% was obtained at the cut-off level 400. Conclusions: RMI have a significant predictability in differentiating benign and malignant adnexal masses, thus can effectively be used in clinical practice.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.59 no.10
• /
• pp.1882-1888
• /
• 2010

It is widely known that ESA(Electric Signature Analysis) method is very useful one for fault diagnosis of an induction motor. Online fault diagnosis system of induction motors using LabVIEW is proposed to detect the fault of broken rotor bars and shorted turns in stator. This system is not model-based system of induction motor but LabVIEW-based fault diagnosis system using FFT spectrum of stator current in faulty motor without estimating of motor parameters. FFT of stator current in faulty induction motor is measured and compared with various reference fault data in data base to diagnose the fault. This paper is focused on to predict and diagnose of the health state of induction motors in steady state. Also, it can be given to motor operator and maintenance team in order to enhance an availability and maintainability of induction motors. Experimental results are demonstrated that the proposed system is very useful to diagnose the fault and to implement the predictive maintenance of induction motors.

• Transactions of The Korea Fluid Power Systems Society
• /
• v.1 no.2
• /
• pp.1-7
• /
• 2004

In-process diagnosis is essential to achieve predictive maintenance in industrial plants. An in- process diagnosis method was proposed for hydraulic servo systems, which was based upon leakage flow measurement. Leakage due to servovalve wear was analysed and modeled mathematically far computer simulation work. The key idea of diagnosis algorithm is that when monitoring signals, such as servovalve input and load displacement are in steady states, the return-line flow of hydraulic servo systems can be regarded as null-leakage of servovalve. Virtual experiments were performed to ensure effectiveness of the proposed diagnosis method.

• The Journal of the Korea Contents Association
• /
• v.11 no.10
• /
• pp.342-348
• /
• 2011

The study aimed to compare the validity between the abdominal ultrasonographic(US) grading system of fatty liver and histologic grading system of fatty liver in living liver donor candidates. As the fatty liver is defined as pathologic total fat >10%, US validity was sensitivity 64.6%, specificity 68%, positive predictive value 76.8%, negative predictive value 54%. As the strict data handling on US grading normal, mild fatty liver are negative, moderate fatty liver is positive, US validity was sensitivity 26.8%, specificity 100%, positive predictive value 100%, negative predictive value 45.5%. ROC curve analysis according to different cut off value of liver-to-kidney brightness ratio was Area under ROC curve=0.859(95% CI=0.795~0.922, state variable= total fat 10%). There were statistically significant difference( p<0.001). Ultrasonography for the fatty diagnosis showed a high validity to predict the result of histology grade of fatty liver.

• Asian Pacific Journal of Cancer Prevention
• /
• v.13 no.8
• /
• pp.3611-3615
• /
• 2012

Objective: The objective of this study was to establish the diagnostic accuracy, specificity and sensitivity of fine needle aspiration cytology(FNAC) for intra-oral tumors, comparing with histopathology as the gold standard. Materials and methods: Forty cases of FNA cytology from intraoral tumors was performed in AFID along with the demographic data and clinical information and then diagnosed at AFIP, Rawalpindi. Then the cytology results obtained per FNAC were compared with the histopathological biopsy results of the same lesions. The following variables were recorded for each patient: Age, gender, site of biopsy, diagnosis. The data were entered and analyzed using Open-epi version 2.0. Diagnostic accuracy, sensitivity, specificity, positive predictive value and negative predictive value were calculated. Cohen Kappa was further applied to compare the agreement between the biopsy and FNAC diagnoses. A p-value of < 0.05 was considered as statistically significant. Results: Among the total patients included in the study there were 24 males and 16 females, with a ratio of 1.5:1. Age of the patients ranged from 24 to 80 years with a mean of 52 years. A total of six sites were aspirated from the oral cavity with maximum (11) aspirates taken from alveolar ridge. The results of FNAC revealed that there were 32 malignant and 8 benign aspirates. Confirmation through histopathological analysis came for 31/32 malignant cases while one was falsely given positive for malignancy on FNAC. Among a total of 40 cases, 31(77%) cases diagnosed were found to be malignant and remaining 9(23%) were benign. The FNAC results revealed 32 malignant and 8 benign lesions. Histopathology of the subsequent surgically excised specimen showed malignant lesions in 31(77%) and benign in 9(23%) patients. As a whole, it was found that the absolute sensitivity for introral FNAC was 100% and specificity 89% with positive predictive value of 97% and negative predictive value of 100%. Conclusion: Cytological diagnosis was almost corroborative with final histopathological diagnosis in all cases, with very few exceptions, exhibiting high diagnostic accuracy.

• Asian Pacific Journal of Cancer Prevention
• /
• v.16 no.2
• /
• pp.707-712
• /
• 2015

Background: To evaluate use of magnetic resonance imaging (MRI) and a logistic model including risk factors for lymph node metastasis for improved diagnosis. Materials and Methods: The subjects were 176 patients with rectal cancer who underwent preoperative MRI. The longest lymph node diameter was measured and a cut-off value for positive lymph node metastasis was established based on a receiver operating characteristic (ROC) curve. A logistic model was constructed based on MRI findings and risk factors for lymph node metastasis extracted from logistic-regression analysis. The diagnostic capabilities of MRI alone and those of the logistic model were compared using the area under the curve (AUC) of the ROC curve. Results: The cut-off value was a diameter of 5.47 mm. Diagnosis using MRI had an accuracy of 65.9%, sensitivity 73.5%, specificity 61.3%, positive predictive value (PPV) 62.9%, and negative predictive value (NPV) 72.2% [AUC: 0.6739 (95%CI: 0.6016-0.7388)]. Age (<59) (p=0.0163), pT (T3+T4) (p=0.0001), and BMI (<23.5) (p=0.0003) were extracted as independent risk factors for lymph node metastasis. Diagnosis using MRI with the logistic model had an accuracy of 75.0%, sensitivity 72.3%, specificity 77.4%, PPV 74.1%, and NPV 75.8% [AUC: 0.7853 (95%CI: 0.7098-0.8454)], showing a significantly improved diagnostic capacity using the logistic model (p=0.0002). Conclusions: A logistic model including risk factors for lymph node metastasis can improve the accuracy of MRI diagnosis of rectal cancer.

• Journal of Chest Surgery
• /
• v.33 no.8
• /
• pp.669-675
• /
• 2000

Background: Tuberculous pleurisy is the leading cause of pleural effusion in Korea. And differential diagnosis of tuberculous pleurisy with other cause is clinically very important. Traditional diagnostic methods such as routine analysis of pleural fluid, staining for acid-fast bacilli or pleural biopsy have major inherent limitaion. This study was designed to evaluate the significance of pleural fluid polymerase chain reaction(PCR) and adenosine deaminase (ADA) activity in early diagnosis of tuberculous pleurisy. Material and Method: Between March 1996 and July 1997, 198 patients with pleural effusion reviewed retrospectively. The study group included 112 cases with tuberculous effusion and 86 cases with non-tuberculous effusions, whose diagnoses were confirmed by pleural biopsy, microbiological methods, or cytology. We compared the results of PCR and pleural fluid levels of ADA between tuberculous and non-tuberculous effusions. Result: Mean age was 47.54$\pm$19.52 years(range 2 to 85 years). The positive rate of PCR was significantly higher in tuberculous group than non-tuberculous group(p<0.05). The sensitivty, specificity, positive predictive value(PPV), and negative predictive value(NPV) for PCR were 31.7, 90.9, 83.0, and 48.8%, respectively. Mean ADA activity was significantly higher in tuberculous group than non-tuberculous group(83.2 U/L vs 49.8 U/L)(p<0.05). With diagnostic thresholds of 40 U/L, the sensitivity, specificity, PPV, and NPV of ADA for tuberculosis were 75.9, 70.9, 77.3, and 69.3% respectively. At a level of 70 U/L, the sensitivity, specificity, PPV, and NPV of ADA for tuberculosis were 70.1, 75.9, 82.9, and 60.3% respectively. Conclusion: PCR is very highly specific, but less sensitive methods in diagnosis of tuberculous pleurisy. But ADA level of pleural fluid has acceptable sensitivity and specificity in diagnosis of tuberculous pleurisy. ADA activity is more useful test in the evaluation of pleural effusions.

• Journal of Electrical Engineering and Technology
• /
• v.12 no.2
• /
• pp.886-889
• /
• 2017

An active Fault Tolerant Model Predictive Control (FTMPC) using Fuzzy scheduler is developed. Fault tolerant Control (FTC) system stages are broadly classified into two namely Fault Detection and Isolation (FDI) and fault accommodation. Basically, the faults are identified by means of state estimation techniques. Then using the decision based approach it is isolated. This is usually performed using soft computing techniques. Fuzzy Decision Making (FDM) system classifies the faults. After identification and classification of the faults, the model is selected by using the information obtained from FDI. Then this model is fed into FTC in the form of MPC scheme by Takagi-Sugeno Fuzzy scheduler. The Fault tolerance is performed by switching the appropriate model for each identified faults. Thus by incorporating the fuzzy scheduled based FTC it becomes more efficient. The system will be thereafter able to detect the faults, isolate it and also able to accommodate the faults in the sensors and actuators of the Continuous Stirred Tank Reactor (CSTR) process while the conventional MPC does not have the ability to perform it.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.35 no.4
• /
• pp.10-15
• /
• 2012

The Mahalanobis-Taguchi System is a diagnosis and predictive method for analyzing patterns in multivariate cases. The goal of this study is diagnosis of the spondylolisthesis from biomedical data that is derived from the shape and orientation of the pelvis and lumbar spine. The data set has six attributes including pelvic incidence, pelvic tilt, lumbar lordosis angle, sacral slope, pelvic radius and grade of spondylolisthesis and two class including normal and abnormal. From University of California at Irvine machine learning repository, 100 normal and 150 spondylolisthesis patient's data were used for this study. Mahalanobis Taguchi System (MTS) application process and the diagnosis results were described in this paper.