• 치위생과학회지
• /
• 제15권3호
• /
• pp.377-382
• /
• 2015

본 연구는 노인 64명을 대상으로 Qraycam을 이용한 이미지로 치면세균막 지수를 측정하여 착색 검사와의 신뢰도를 평가하고자 하였으며, 수집된 자료를 분석하여 다음과 같은 결과를 얻었다. 측정 방법, 측정 부위, 지수유형별에 따른 치면세균막 지수의 각 측정값 사이에 높은 일치도가 나타났다. 측정 방법에 따른 전치부 순면의 Quigley-Hein index와 PCR의 평균은 유의한 차이가 없었으며, 측정 부위에 따른 Quigley-Hein index와 PCR의 평균은 통계적으로 유의한 차이가 있었다(p<0.001). Qraycam 검사와 착색 검사의 지수유형별 kappa계수를 확인한 결과 kappa값의 평균은 Quigley-Hein index 0.90, PCR 0.84로 전체적으로 높은 일치도를 보였다. Qraycam과 착색 검사의 전치부 순면 및 전체 치아를 Quigley-Hein index와 PCR로 측정하였을 때, 변수 간 ICC는 1에 가까운 결과를 보였다. 본 연구 결과 측정방법, 측정 부위, 지수유형별에 따른 치면세균막 지수의 일치도 평가에서 모든 측정값의 일치도가 높게 나타났다. 따라서 Qraycam은 치면세균막 검사 시 screening 도구로서 충분한 신뢰도가 있음을 확인하였다.

• 대한원격탐사학회지
• /
• 제35권6_1호
• /
• pp.999-1009
• /
• 2019

북극항로의 개척 가능성과 정확한 기후 예측 모델의 필요성에 의해 북극해 고해상도 해빙 지도의 중요성이 증가하고 있다. 그러나 기존의 북극 해빙 지도는 제작에 사용된 위성 영상 취득 센서의 특성에 따른 데이터의 취득과 공간해상도 등에서 그 활용도가 제한된다. 본 연구에서는 Sentinel-1 A/B SAR 위성자료로부터 고해상도 해빙 지도를 생성하기 위한 딥러닝 기반의 해빙 분류 알고리즘을 연구하였다. 북극해 Ice Chart를 기반으로 전문가 판독에 의해 Open Water, First Year Ice, Multi Year Ice의 세 클래스로 구성된 훈련자료를 구축하였으며, Convolutional Neural Network 기반의 두 가지 딥러닝 모델(Simple CNN, Resnet50)과 입사각 및 thermal noise가 보정된 HV 밴드를 포함하는 다섯 가지 입력 밴드 조합을 이용하여 총 10가지 케이스의 해빙 분류를 실시하였다. 이 케이스들에 대하여 Ground Truth Point를 사용하여 정확도를 비교하고, 가장 높은 정확도가 나온 케이스에 대해 confusion matrix 및 Cohen의 kappa 분석을 실시하였다. 또한 전통적으로 분류를 위해 많이 활용되어 온 Maximum Likelihood Classifier 기법을 이용한 분류결과에 대해서도 같은 비교를 하였다. 그 결과 Convolution 층 2개, Max Pooling 층 2개를 가진 구조의 Convolutional Neural Network에 [HV, 입사각] 밴드를 넣은 딥러닝 알고리즘의 분류 결과가 96.66%의 가장 높은 분류 정확도를 보였으며, Cohen의 kappa 계수는 0.9499로 나타나 딥러닝에 의한 해빙 분류는 비교적 높은 분류 결과를 보였다. 또한 모든 딥러닝 케이스는 Maximum Likelihood Classifier 기법에 비해 높은 분류 정확도를 보였다.

• 한국지리정보학회지
• /
• 제27권1호
• /
• pp.115-127
• /
• 2024

본 연구는 Transformer 모듈을 기반으로 다양한 구조의 모델을 구성하고, 토지피복 분류를 수행하여 Transformer 모듈의 활용방안 검토를 목적으로 하였다. 토지피복 분류를 위한 딥러닝 모델은 CNN 구조를 가진 Unet 모델을 베이스 모델로 선정하였으며, 모델의 인코더 및 디코더 부분을 Transformer 모듈과 조합하여 총 4가지 딥러닝 모델을 구축하였다. 딥러닝 모델의 학습과정에서 일반화 성능 평가를 위해 같은 학습조건으로 10회 반복하여 학습을 진행하였다. 딥러닝 모델의 분류 정확도 평가결과, 모델의 인코더 및 디코더 구조 모두 Transformer 모듈을 활용한 D모델이 전체 정확도 평균 약 89.4%, Kappa 평균 약 73.2%로 가장 높은 정확도를 보였다. 학습 소요시간 측면에서는 CNN 기반의 모델이 가장 효율적이었으나 Transformer 기반의 모델을 활용할 경우, 분류 정확도가 Kappa 기준 평균 0.5% 개선되었다. 차후, CNN 모델과 Transformer의 결합과정에서 하이퍼파라미터 조절과 이미지 패치사이즈 조절 등 다양한 변수들을 고려하여 모델을 고도화 할 필요가 있다고 판단된다. 토지피복 분류과정에서 모든 모델이 공통적으로 발생한 문제점은 소규모 객체들의 탐지가 어려운 점이었다. 이러한 오분류 현상의 개선을 위해서는 고해상도 입력자료의 활용방안 검토와 함께 지형 정보 및 질감 정보를 포함한 다차원적 데이터 통합이 필요할 것으로 판단된다.

• Tuberculosis and Respiratory Diseases
• /
• 제79권4호
• /
• pp.282-288
• /
• 2016

Background: Tuberculosis (TB) is a major health problem, and accurate and rapid diagnosis of multidrug-resistant (MDR) and extended drug-resistant (XDR) TB is important for appropriate treatment. In this study, performances of solid and liquid culture methods were compared with respect to MDR- and XDR-TB isolate recovery and drug susceptibility testing. Methods: Sputum specimens from 304 patients were stained with Ziehl-Neelsen method. Mycobacterium tuberculosis (Mtb) isolates were tested for recovery on $L{\ddot{o}wenstein$-Jensen (LJ) medium and the BacT Alert 3D system. For drug susceptibility testing of Mtb, isolates were evaluated on M-KIT plates and the BacT Alert 3D system. Results: The recovery rates were 94.9% (206/217) and 98.2% (213/217) for LJ medium and the BacT Alert 3D system, respectively (kappa coefficient, 0.884). The rate of drug resistance was 13.4% for at least one or more drugs, 6.0% for MDR-TB and 2.3% for XDR-TB. M-KIT plate and BacT 3D Alert 3D system were comparable in drug susceptibility testing for isoniazid (97.7%; kappa coefficient, 0.905) and rifampin (98.6%; kappa coefficient, 0.907). Antibiotic resistance was observed using M-KIT plates for 24 of the total 29 Mtb isolates (82.8%). Conclusion: The liquid culture system showed greater reduction in the culture period, as compared with LJ medium; however, drug susceptibility testing using M-KIT plates was advantageous for simultaneous testing against multiple drug targets.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2003년도 춘계학술대회
• /
• pp.2168-2173
• /
• 2003

Pressure balancing valve is one of important control devices, which is fully automatic and no manual controls, regulating or adjustments are needed. It is typically used to maintain constant temperature of working fluid in power and chemical plants and domestic water supply systems. Pressure balancing valve is composed of body, cylinder and balancing piston. Therefore, the balancing piston shapes are important design parameters for a pressure balancing valve. In this study, numerical and experimental analyses are carried out with two different balancing piston shapes. Especially, the distribution of static pressure is investigated to calculate the flow coefficient($C_v$). The governing equations are derived from making using of three-dimensional Navier-Stokes equations with standard ${\kappa}-{\varepsilon}$ turbulence model and SIMPLE algorithm. Using commercial code, PHOEIC, the pressure and flow fields in pressure balancing valve are depicted.

• Imaging Science in Dentistry
• /
• 제47권1호
• /
• pp.25-31
• /
• 2017

Purpose: The aim of this study was to compare the diagnostic accuracy of previously trained endodontists in the detection of artificially created periapical lesions using cone-beam computed tomography (CBCT) and digital periapical radiography (DPR). Materials and Methods: An ex vivo model using dry skulls was used, in which simulated apical lesions were created and then progressively enlarged using #1/2, #2, #4, and #6 round burs. A total of 11 teeth were included in the study, and 110 images were obtained with CBCT and with an intraoral digital periapical radiographic sensor (Instrumentarium dental, Tuusula, Finland) initially and after each bur was used. Specificity and sensitivity were calculated. All images were evaluated by 10 previously trained, certified endodontists. Agreement was calculated using the kappa coefficient. The accuracy of each method in detecting apical lesions was calculated using the chisquare test. Results: The kappa coefficient between examiners showed low agreement (range, 0.17-0.64). No statistical difference was found between CBCT and DPR in teeth without apical lesions (P=.15). The accuracy for CBCT was significantly higher than for DPR in all corresponding simulated lesions(P<.001). The correct diagnostic rate for CBCT ranged between 56.9% and 73.6%. The greatest difference between CBCT and DPR was seen in the maxillary teeth (CBCT, 71.4%; DPR, 28.6%; P<.01) and multi-rooted teeth (CBCT, 83.3%; DPR, 33.3%; P<.01). Conclusion: CBCT allowed higher accuracy than DPR in detecting simulated lesions for all simulated lesions tested. Endodontists need to be properly trained in interpreting CBCT scans to achieve higher diagnostic accuracy.

• 한국산림과학회지
• /
• 제81권4호
• /
• pp.346-356
• /
• 1992

미서부(美西部)의 광대한 Snake강(江) 범람평원은 홍수로 인하여 수로(水路) 및 식생형(植生型)의 빈번한 변화 및 침해를 받아왔다. 1985년과 1988년 기간 동안의 습지대 식생형의 변화를 탐지하기 위하여, 원격탐사의 변화탐지 기법(技法) 중 구분후(區分后) 비교(比較) 및 영상대차법(映像對差法) 등을 Landsat-5호 TM 디지탈 데이타를 이용하여 비교 고찰 하였다. 대차(對差)된 적외선대(外線帶) 영상들이 가시대(可視帶) 영상을보다 나은 정확도 지표(指標)를 보였으며, 역기법(閾技法)을 적용하여, 영상대차법에 의하여 변형된 영상들로부터 변화(變化)와 무변화(無變化)를 구분하였다. 또한, 여러 정확도 지표들 즉, 카파 일치계수(一致係數), 총정확도, 생산자 정확도, 이용자 정확도 및 평균정확도(생산자 및 이용자 정확도 등에 근거한) 등을 이용하여 최적역영역(最適閾領域)을 결정함에 있어서의 문제점들을 고찰하였다.

• 동의생리병리학회지
• /
• 제32권3호
• /
• pp.171-177
• /
• 2018

The purpose of this study was to evaluate the reliability and the validity of Deficiency and Excess Pattern Identification Questionnaire. The number of subjects enrolled in this study was 248, from July 2015 to March. 2016. The surveys was conducted two times with 3 month interval. The Cronbach's ${\alpha}$ analysis for internal reliability, Pearson Correlation Coefficient analysis for test-retest reliability were conducted. Factor analysis with varimax rotation for construct validity was used. Kappa analysis for diagnostic reliability were used. The significant p-value was < .05. The Cronbach's ${\alpha}$ was .929 in Deficiency and .932 in Excess questionnaire. The reliabilities between test and retest Intra Correlation Coefficient (ICC) was .71-.762 in the Deficiency, and .58-.786 in Excess questionnaire, respectively. Deficiency was divided by five factors, and Excess four factors. The factor convergence was 72.54% in the Deficiency and 67.5% in Excess questionnaire. The test-retest agreement of four pattern was 68.5% and Kappa was .530. This study reveals that Deficiency and Excess Pattern Identification Questionnaire is a reliable and valid. However, further study to validate the questionnaire is needed.

• Asian Pacific Journal of Cancer Prevention
• /
• 제17권2호
• /
• pp.491-495
• /
• 2016

Background: Small B-cell non-Hodgkins lymphoma (NHL) is difficult to be distinguished from non-neoplastic reactive processes using conventional haematoxylin-eosin (HE) staining due to different interpretations among pathologists with diagnosis based on morphologic features. Ancillary examinations such as immunohistochemical (IHC) staining are essential. However, negative or doubtful results are still sometimes obtained due to unsatisfactory tissue processing or IHC technique. The polymerase chain reaction (PCR) as a molecular diagnostic technique is very sensitive and specific. Clonality detection of heavy chain immunoglobulin (IgH) gene rearrangement has been widely used to establish diagnosis of B-cell NHL. Aims: To elaborate interobserver variation in small B-cell NHL diagnosis based on morphologic features only and to confirm sensitivity and specificity of the PCR technique as an ancillary method. Materials and Methods: A toptal of 28 samples of small B cell NHL and suspicious lymphoma were interpreted by 3 pathologists in Sardjito General Hospital based on their morphology only. The reliability of assessment and the coefficient of interobserver agreement were calculated by Fleiss kappa statistics. Interpretation results were confirmed with IHC staining (CD20, CD3, Bcl2). PCR was performed to analyze the clonality of IgH gene rearrangement. Results: Interobserver agreement in morphologic evalution of small B cell NHL and chronic lymphadenitis revealed kappa coefficient 0.69 included in the substantial agreement category. The cases were divided into 3 groups based on morphology and IHC results; lymphoma, reactive process and undetermined group. PCR analysis showed 90% sensitivity and 60% specificity. Conclusions: The present study revealed a substantial agreement among pathologists in small B-cell NHL diagnosis. For difficult cases, PCR is useful as complementary method to morphologic and IHC examinations to establish definitive diagnosis.

• 아동학회지
• /
• 제26권2호
• /
• pp.27-41
• /
• 2005

The purpose of this study was to test the reliability and validity of the Infant Montessori Performance Assessment Scales developed by the authors. Subjects of the study consisted of 132 Montessori teachers and 50 infants who were attending Montessori home school throughout the whole country, using census method. Infant's age ranged from 6 to 24 months. Data were analysed by using SAS 8.2 PC program. Statistical methods employed were frequency, Cronbach's alpha, Kappa coefficient, test-retest correlation, construct validity, and concurrent validity. The Cronbach's alpha of 6 sub-scales included physical, creativity, practical life, sense, language, and cognitive education, which ranged from .70 to .86. And the! correlation of the test/re-test was .72. The correlation between the 6 sub-scales of Infant-Montessori Performance Assessment Scales and the total scores of 6 items ranged from .06 to .84. Therefore, the construct validity of Infant Montessori Performance Assessment Scales was verified. The Kappa coefficient of inter-rater reliability was .76. The correlation between the Infant Montessori Performance Assessment Scales and the Standardized Korean Creativity Traits Checklist(K-CTC) and the Korean Child Social Maturity Scales showed non-significant levels of .16 and .12 respectively. Conclusively, Infant Montessori Performance Assessment Scales developed by the authors were verified through the above reliability and validity tests. Specifically the Infant Montessori Performance Assessment Scales showed the relationship of the convergent and divergent validity with the Korean Child Social Maturity Scales and the Standardized Korean Creativity Traits Checklist, respectively.