• 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.

• Convergence Security Journal
• /
• v.23 no.5
• /
• pp.65-72
• /
• 2023

This paper proposes a method for detecting malicious domains considering human habitual characteristics by building a Deep Learning model based on LSTM (Long Short-Term Memory). DGA (Domain Generation Algorithm) malicious domains exploit human habitual errors, resulting in severe security threats. The objective is to swiftly and accurately respond to changes in malicious domains and their evasion techniques through typosquatting to minimize security threats. The LSTM-based Deep Learning model automatically analyzes and categorizes generated domains as malicious or benign based on malware-specific features. As a result of evaluating the model's performance based on ROC curve and AUC accuracy, it demonstrated 99.21% superior detection accuracy. Not only can this model detect malicious domains in real-time, but it also holds potential applications across various cyber security domains. This paper proposes and explores a novel approach aimed at safeguarding users and fostering a secure cyber environment against cyber attacks.

• Clinical and Experimental Reproductive Medicine
• /
• v.36 no.4
• /
• pp.255-263
• /
• 2009

Objective: The objective of the study was to determine the diagnostic performance of transabdominal ultrasound by receiver operating characteristic (ROC) curve analysis, in order to evaluate the usefulness in establishing the diagnosis of polycystic ovary syndrome (PCOS). Methods: Questionnaires were given to 8,793 reproductive women reviewed at Ewha Womans University Mokdong hospital. Ultrasound examinations were performed in 701 women with a transabdominal transducer. Transabdominal ultrasounds were performed in 185 normal control women (normal menstruation without hyperandrogenism or PCO morphology) and 248 PCOS patients according to National Institutes of Health (NIH) PCOS diagnosis criteria. ROC curves were calculated for ovarian volume and follicle number. Results: In normal control group, the mean age were $23.64{\pm}4.26$ years old and the mean ovarian volume and follicle number were $6.03{\pm}1.89\;cm^3$ and $6.49{\pm}1.93$, respectively. The ovarian volume showed an area under the ROC curve (AURC) of 0.761. A ovarian volume decision threshold >$9\;cm^3$ had a sensitivity of 51.0% and a specificity of 91.4% for the diagnosis of PCOS. The follicle number showed an AURC of 0.733. A follicle number decision threshold ${\geq}9$ had a sensitivity of 54.9% and a specificity of 87.0% for the diagnosis of PCOS. A follicle number decision threshold ${\geq}10$ had a sensitivity of 53.2% and a specificity of 90.4%. A follicle number and a ovarian volume did not have a high diagnostic power for screening for PCOS. Conclusion: Our results suggest that transabdominal ultrasound assessment is not effective for the detection of PCOS in young women of reproductive age.

• International Journal of Highway Engineering
• /
• v.20 no.2
• /
• pp.35-44
• /
• 2018

PURPOSES : The purpose of this research is to analyze the characteristics of panels that affect the evaluating results of riding quality and to evaluate the appropriateness of roughness management criteria based on ride comfort satisfaction. METHODS : In order to analyze the influence of panel characteristics of riding quality, 33 panels, consisting of civilians and experts, were selected. Also, considering the roughness distribution of the expressway, 35 sections with MRI ranging from 1.17 m/km to 4.65 m/km were selected. Each panel boarded a passenger car and evaluated the riding quality with grades from 0 to 10, and assessed whether it was satisfied or not. After removing outlier results using a box plot technique, 964 results were analyzed. An ANOVA was conducted to evaluate the effects of panel expertise, age, driving experience, vehicle ownership, and gender on the evaluation results. In addition, by using the receiver operating characteristics (ROC) curve, the MRI value, which can most accurately evaluate the satisfaction with riding quality, was derived. Then, the compatibility of MRI was evaluated using AUC as a criterion to assess whether the riding quality was satisfactory. RESULTS : Only the age of the panel participants were found to have an effect on the riding quality satisfaction. It was found that satisfaction with riding quality and MRI are strongly correlated. The satisfaction rate of roughness management criteria on new (MRI 1.6 m/km) and maintenance (MRI 3.0 m/km) expressways were 95% and 53%, respectively. As a result of evaluating the roughness management criteria by using the ROC curve, it was found that the accuracy of satisfaction was the highest at MRI 3.1-3.2 m/km. In addition, the AUC of the MRI was about 0.8, indicating that the MRI was an appropriate index for evaluating the riding quality satisfaction. CONCLUSIONS : Based on the results, the distribution of the panels' age should be considered when panel rating is conducted. From the results of the ROC curve, MRI of 3.0 m/km, which is a criterion of roughness management on maintenance expressways, is considered as appropriate.

• Korean Journal of Clinical Pharmacy
• /
• v.20 no.2
• /
• pp.107-113
• /
• 2010

서 론: 중환자실에서 집중적 인슐린 요법에 의한 평균혈당강하는 사망률을 감소시키는 것으로 나타났으나 이로 인한 저혈당 및 혈당변동은 새로운 문제로 대두되고 있다. 본 연구에서는 혈당과 관련한 여러 요인들이 사망에 미치는 영향을 규명하고 적정 혈당치를 확인하고자 하였다. 연구방법: 2008년 2월부터 7월 사이인 6개월 동안 서울아산병원 외과계 중환자실에서 4일 이상 재실한 18세 이상의 성인 환자를 대상으로 전자 의무기록 조사를 통해 후향적으로 연구가 진행되었다. 연구를 위해 환자의 인구학적 특성, 수술의 종류, 중환자실에서의 재실기간, 사망여부, 스테로이드 사용 유무, 기계적 인공호흡기의 사용유무, 신대체요법의 사용유무, 혈당치, 재실기간 중 스테로이드 사용유무와 인슐린 양, 입원 후 첫 24시간 동안의 포도당 주입속도, 입원 후 2일 이내와 그 이후에 발생한 균혈증 감염, APACH II와 SOFA 점수를 조사하였다. 혈당수치는 각각의 환자에서 중환자실 입실 후 가장 처음 측정된 혈당, 재실기간 중 가장 높은 혈당과 가장 낮은 혈당수치를 조사하였고 중환자실 전체 재실기간 동안 혈당수치의 평균과 변동계수를 계산하였다. 이상의 혈당관련지표를 포함한 인자들이 일차 종속변수인 사망에 어떠한 영향을 주는지를 환자를 생존군과 사망군으로 나누어 분석하였고 ROC (receiver operator characteristic) 곡선을 사용하여 혈당지표와 APACH 및 SOFA 점수의 cut-off치를 구하여 이로부터 단변량 및 다변량 분석을 시행하였다. 결 과: 연구에 포함된 환자는 170명 이었고 그 중 23명이 연구 기간 중 중환자실에서 사망하였다. 생존자에 비해 사망자의 최대혈당은 유의적으로 높았고 최소혈당치는 유의적으로 낮아 높은 변동계수를 보였다. ROC곡선으로부터 산출된 혈당치들의 cut-off 수치는 최소혈당치 70 mg/dL, 변동계수 25%, 최대혈당치 250 mg/dL, 평균혈당치 150 mg/dL이었다. 다변량분석에서 최소혈당이 70 mg/dL 보다 큰 경우가 낮은 경우에 비해 오즈비가 0.922(95% 신뢰구간 0.881-0.965)로 유의성 있게 낮았으며 변동계수가 25% 보다 높은 집단의 경우 그보다 낮은 집단에 비해 오즈비가 1.121(95% 신뢰구간 1.017-1.236)로 유의성 있게 높았다. Kaplan-Meier 생존분석 결과 최소혈당치 70 mg/dL와 변동계수 25%에 따라 생존기간에 유의성 있는 차이가 나타났다.(각각 P < 0.001, P < 0.05) 결 론: 고혈당 발생의 감소뿐 아니라 최소혈당치를 70 mg/dL 이상으로 유지하면서 변동을 최소화하는 것이 외과계 중환자실에서의 사망률감소를 위한 중요한 요인임을 알 수 있었다.

• Journal of Veterinary Clinics
• /
• v.29 no.1
• /
• pp.18-22
• /
• 2012

This study was performed to evaluate the sensitivity of conventional film-screen radiography (CFSR) and direct digital radiography (DDR) for detection of various amounts of free peritoneal fluid. Ten adult male healthy beagles were used in this study. Radiographic examinations were performed in the right lateral and ventrodorsal positions. Fluid was injected in increments of 2.0 ml/kg of body weight up to 20.0 ml/kg of body weight. The images of CFSR and DDR were evaluated by two veterinary radiologists for evidence of abdominal fluid without knowledge of injected fluid volume. Data were evaluated by using the receiver operation curve (ROC) analysis and the area under the curve (AUC). There was no significant difference in detection of peritoneal fluid between DDR and CFSR in the ROC analysis. The accuracy of CFSR (0.805) was relatively higher than that of DDR (0.733), based on the ROC analysis and AUC. AUC of CFSR was higher in most injection doses. These results suggest that CFSR is more accurate than DDR for the detecting peritoneal fluid. Therefore, for situation in which digital radiographs are equivocal or small amount of fluid is suspected, other imaging modalities, such as ultrasonography would be helpful for determining the presence of fluids.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.36 no.2
• /
• pp.51-58
• /
• 2018

Recently, various satellite sensors have been developed and it is becoming more convenient to acquire multitemporal satellite images. Therefore, various researches are being actively carried out in the field of utilizing change detection techniques such as disaster and land monitoring using multitemporal satellite images. In particular, researches related to the development of unsupervised change detection techniques capable of extracting rapidly change regions have been conducted. However, there is a disadvantage that false detection occurs due to a spectral difference such as a seasonal change. In order to overcome the disadvantages, this study aimed to reduce the false alarm detection due to seasonal effects using the direction vector generated by applying the $S^2CVA$ (Sequential Spectral Change Vector Analysis) technique, which is one of the unsupervised change detection methods. $S^2CVA$ technique was applied to RapidEye images of the same and different seasons. We analyzed whether the change direction vector of $S^2CVA$ can remove false positives due to seasonal effects. For the quantitative evaluation, the ROC (Receiver Operating Characteristic) curve and the AUC (Area Under Curve) value were calculated for the change detection results and it was confirmed that the change detection performance was improved compared with the change detection method using only the change magnitude vector.

• Journal of the Korean Society of Radiology
• /
• v.13 no.2
• /
• pp.209-216
• /
• 2019

The purpose of this study was to suggest the optimum ambient illumination level for proper visualization in image inspection and reading on CRT and LCD monitors used for ultrasound and reading. The evaluators were divided into 4 groups: 20 (Ultra-sonographer: 20 groups (4 groups: ultra-sonographer, 1-5 years, 5 ultra-sonographers, 6 to 10 years, 5 ultra-sonographers, 11 to 15 years, The subjects were 32 questions. The evaluation method was image evaluation of ultrasonic soft copy images for 30 seconds per 10, 25, 100Lux ambient illumination. The evaluation results were evaluated as 6 points (Normal = Definitely no lesion), 2 points = possibly not a lesion, 3 points = probably not a lesion, 4 points = possibly a lesion, 5 points = probably a lesion, 6 points = Definitely a lesion). In this study, the results of ROC analysis according to ambient light illumination reading softcopy images used for lesion detection of all ultrasound images showed the highest sensitivity, specificity, and AUC results at 10Lux. It was found that optimal use of 10Lux for ambient light illumination would provide optimal detection of lesions in ultrasound soft copy images. Based on the future research data, it will be presented as basic data for designing ambient light brightness of ultrasound imaging laboratory and reading room.

• The Korean Journal of Nuclear Medicine Technology
• /
• v.22 no.1
• /
• pp.98-102
• /
• 2018

Purpose When examining patients with digestive system Leakage, it is not easy to distinguish between bile leakage and bleeding only with Planar images. I would like to evaluate the utility of leakage of bile, confirmation of gastrointestinal bleeding and location discrimination using SPECT-CT. Materials and Methods SPECT-CT was performed according to the request of the reading room after acquiring Planar image for a total of 13 patients, including 8 patients suspected of bile leakage and 5 patients suspected of gastrointestinal bleeding, among patients visiting this specification went. We used Symbia T16 from SIEMENS and Discovery 670 from GE. Planar and SPECT-CT images were evaluated with a score of 1 to 10 by 4 experienced nuclear medicine doctors. Using the sensitivity and the specificity, the evaluation of the inspection by the ROC curve was carried out. The final diagnosis was confirmed by follow-up observation as a result of reoperative surgery. Results The sensitivity, specificity and accuracy of SPECT-CT were 91.7%, 100% and 94.2%, respectively. Planar images were 83.3%, 68.8% and 78.8%, respectively. Planar images showed lower diagnostic accuracy compared to SPECT-CT (78.8%, 94.2%, p<0.05). Moreover, the reliability of the diagnosis of SPECT-CT by ROC curve analysis showed a more useful result than the Planar image (p<0.05). Conclusion SPECT-CT had high sensitivity and specificity for diagnosis of biliary leakage and gastrointestinal bleeding and location discrimination. When SPECT-CT is additionally performed together with Planar images, it may be considered to improve bile leakage and diagnosis of gastrointestinal bleeding patients and position discrimination evaluation. There is usefulness depending on the patient's age and position, but consideration of additional CT exposure may be done sufficiently.

• Nuclear Medicine and Molecular Imaging
• /
• v.41 no.6
• /
• pp.538-545
• /
• 2007

Purpose: We evaluated correlation of $^{18}F$-FDG uptakes, therapeutic response and relapse in pre-treatment $^{18}F$-FDG PET/CT in patients with SCLC. Materials and methods: We included 26 patients with pathologically proven small cell lung cancer. Total 102 lesions (26 lungs, 69 lymph nodes and 8 metastatic lesions) were evaluated. All patients underwent $^{18}F$-FDG PET/CT for staging. The maxSUV was used as a parameter of $^{18}F$-FDG uptake. The patients were divided into responders and non-responders according to response criteria on chest CT scan after 3 cycles of chemotherapy. We compared maxSUV between two groups by using independent t-test. To access correlation with $^{18}F$-FDG uptake and relapse, maxSUV and interval time to relapse was analyzed by correlation analysis. The cutoff value of maxSUV was evaluated by ROC curve. Results: Twelve-one patients (81%) were responders and five patients were non-responders on follow-up chest CT scan. The mean maxSUV of main lung lesions in responders and non-responders were $14.15{\pm}3.72$ and $9.17{\pm}2.15$, respectively. The maxSUV in the responders was significantly lower than that in non-responders (p<0.05). According to ROC curve, point of cut that predicts therapeutic response was 8.98 with 100% sensitivity and 57% specificity. The correlation analysis between $^{18}F$-FDG uptakes and interval time to relapse showed a significant negative correlation (p<0.05, r=-0.757). Conclusion: The pre-treatment $^{18}F$-FDG uptake of responders was significantly lower than that of non-responders. Patients with high $^{18}F$-FDG uptake in pre-treatment $^{18}F$-FDG PET/CT relapse earlier.