http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
갑상선 기능 저하 환자에서 levothyroxine 투여 시간에 따른 효능 비교: 메타분석
이기표,아영미,최혜덕 한국임상약학회 2020 한국임상약학회지 Vol.30 No.2
Background: Levothyroxine is an essential drug for the treatment of hypothyroidism or related diseases. Several studies have reported an association between the effects of levothyroxine treatment and time of administration, which can be inconsistent. Objective: This study was conducted to compare the levels of thyroid-stimulating hormone or free thyroxine between morning and nighttime dosing of levothyroxine. Methods: We reviewed previously reported relevant articles and conducted a meta-analysis. Results: In total, five studies were included in this meta-analysis. Results showed that thyroid-stimulating hormone (standard difference in means [SE]=0.321; 95% confidence interval [CI], -0.016 to 0.657) and free thyroxine (SE= −1.367; 95% CI, -2.943 to 0.210) levels did not differ significantly between morning (before breakfast) and nighttime (before bedtime) administration. Conclusion: This is the first meta-analysis to evaluate the effects of time of administration on levothyroxine levels in patients with hypothyroidism. Based on our results, we suggest considering patients’ lifestyles or daily routines when counselling them on the optimal time of administration for levothyroxine.
위내시경 영상에서의 위 병변 자동 검출 모델 개발을 위한 RetinaNet 기반 backbone 네트워크에 따른 학습 성능 비교
이기표,김영재,박동균,김재승,김광기 차세대컨버전스정보서비스학회 2023 차세대컨버전스정보서비스기술논문지 Vol.12 No.2
Gastric cancer occurs mostly in Asian countries such as Korea and Japan. Gastroscopy allows diagnosis and treatment of gastric cancer at the same time, and the probability of successful treatment is very high at early detection. However, due to the nature of the inspection which progresses in real time, proficiency and experience of the clinician affect the results, and the accuracy of the inspectioncan decrease due to increased work fatigue and decreased concentration. In this study, we developed a model that automatically detects the regions of gastric lesion in gastroscopic images using the RetinaNet network so that it can be used as an auxiliary system during gastroscopy. We confirmed the detection performance of models trained using ResNet50, ResNet152, EfficientNetB0, and EfficientNetB4 networks in a RetinaNet-based backbone network, and compared the performance between each model. The average sensitivities (FP/images) of RetinaNet-based backbone network-models were 73.72% (0.0489) for ResNet50, 78.26% (0.0458) for ResNet152, 79.67% (0.3268) for EfficinetNetB0, and 79.67% (0.3268) for EfficientNetB4. The EfficientNetB0 network showed the highest sensitivity, but the FP/images were very high, so the network satisfying both performance values was ResNet152. 본 연구에서는 위내시경 검사 시에 보조 시스템으로 활용할 수 있도록 RetinaNet 네트워크를 사용하여 위내시경 영상에서의 위 병변의 위치를 자동으로 검출하는 모델을 개발하였다. 위암은 한국이나 일본 등의 아시아권에서 대부분 발생한다. 그러나 위내시경 검사는 동시에 진단이나 치료할 수 있으며, 조기 발견 시 치료 성공확률이 매우 높다. 그러나 실시간으로 진행되는 검사 특성상 숙련도나 경험이 결과에 영향을 주며, 업무의 피로도 상승과 집중력 하락으로 인해 검사의 정확도가 낮아지게 된다. RetinaNet 기반의 backbone 네트워크로 ResNet50, ResNet152, EfficientNetB0, EfficientNetB4 네트워크를 사용하여 학습한 모델의 검출 성능을 확인하고, 각 모델 간의 성능을 비교하였다. RetinaNet 기반 backbone 네트워크별 모델들의 평균 민감도(FP/images)는 ResNet50 73.72%(0.0489), ResNet152 78.26%(0.0458), EfficinetNetB0 79.67%(0.3268), EfficientNetB4 62.66%(0.0448)를 보였다. EfficientNetB0 네트워크는 가장 높은 민감도를 나타냈으나 FP/images가 매우 높게 나타나 두 성능치를 모두 만족하는 네트워크는 ResNet152였다.
운동심상과 운동결과기대, 운동지속의도의 구조관계에서운동결과기대의 매개효과
이기표,이윤구,윤용진 한국스포츠심리학회 2018 한국스포츠심리학회지 Vol.29 No.2
Purpose: The purpose of this study was to provide fundamental information to evaluate and predict individual's exercise behavior as a psycho-behavioral intervention strategy to maintain exercise participation. This was accomplished by investigating the mediating effect of exercise outcome expectation in the structural relationship of exercise participants' exercise imagery, exercise outcome expectation, and intention to exercise adherence. Method: The study subjects were 303 male and female participating in exercise, and the collected data were analyzed through descriptive statistical analysis, frequency analysis, reliability analysis, confirmatory factor analysis, correlation analysis, structural equation model analysis, Bootstrapping method was used to examine the mediation effect. Results: First, exercise imagery had a positive effect on exercise outcome expectation. Second, exercise imagery had a positive effect on intention to exercise adherence. Third, exercise outcome expectation had a positive effect on intention to exercise adherence. Fourth, exercise outcome expectation partially mediated the relationship between exercise imagery and intention to exercise adherence. Conclusion: This study revealed that imagery, exercise outcome expectation, and intention to exercise adherence were positively correlated, and exercise outcome expectation mediated the relationship between exercise imagery and intention to exercise adherence. However, further research is needed to examine the effect of exercise imagery related to exercise behavior as the exercise imagery research for participants engaged in exercise is still in the early stage in South Korea.
손목 관절 단순 방사선 영상에서 딥 러닝을 이용한 전후방 및 측면 영상 분류와 요골 영역 분할
이기표,김영재,이상림,김광기,Lee, Gi Pyo,Kim, Young Jae,Lee, Sanglim,Kim, Kwang Gi 대한의용생체공학회 2020 의공학회지 Vol.41 No.2
The purpose of this study was to present the models for classifying the wrist X-ray images by types and for segmenting the radius automatically in each image using deep learning and to verify the learned models. The data were a total of 904 wrist X-rays with the distal radius fracture, consisting of 472 anteroposterior (AP) and 432 lateral images. The learning model was the ResNet50 model for AP/lateral image classification, and the U-Net model for segmentation of the radius. In the model for AP/lateral image classification, 100.0% was showed in precision, recall, and F1 score and area under curve (AUC) was 1.0. The model for segmentation of the radius showed an accuracy of 99.46%, a sensitivity of 89.68%, a specificity of 99.72%, and a Dice similarity coefficient of 90.05% in AP images and an accuracy of 99.37%, a sensitivity of 88.65%, a specificity of 99.69%, and a Dice similarity coefficient of 86.05% in lateral images. The model for AP/lateral classification and the segmentation model of the radius learned through deep learning showed favorable performances to expect clinical application.