http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
흉부 X-선 영상에서 밝기값 정규화 및 다중 스케일 폐-집중 패치를 사용한 앙상블 딥러닝 모델 기반의 폐렴 자동 분류
김윤조,안진서,홍헬렌 한국정보과학회 2022 정보과학회논문지 Vol.49 No.9
It is difficult to classify normal and pneumonia in pediatric chest X-ray (CXR) images due to irregular intensity values. In addition, deep learning model has a limitation in that it can misclassify CXR by incorrectly focusing on the outer part of the lung. This study proposed an automatic classification of pneumonia based on ensemble deep learning model using three intensity normalizations and multiscale lung-focused patches on CXR images. First, to correct for irregular intensity values in internal lungs, three intensity normalization methods were performed respectively. Second, to focus on internal lungs, regions of interest were extracted by segmenting lung regions. Third, multiscale lung-focused patches were extracted to train the characterization of pneumonia. Finally, ensemble modeling with attention module was performed to improve the classification performance. In the experiment, the method using large patches of CLAHE images showed an accuracy of 92%, which was 5% higher than that of original images. Furthermore, the proposed method using an ensemble of large and middle patches showed the best performance with an accuracy of 93%. 소아 흉부 X-선 영상(CXR)은 밝기값이 불규칙하여 정상과 폐렴을 구분하기 어렵다. 또한 딥러닝 모델은 폐의 외부 영역에 잘못 집중하여 CXR을 오분류할 수 있다는 한계가 있다. 본 논문은 CXR 영상에서 밝기값 정규화 및 다중 스케일 폐-집중 패치를 사용한 앙상블 딥러닝 기반 폐렴 자동 분류 방법을 제안한다. 첫째, 불규칙한 폐 내부 밝기값을 개선하기 위해 세 가지 밝기값 정규화 방법을 각각 수행한다. 둘째, 폐 내부에 집중하여 학습하기 위해 폐 영역을 분할하여 관심 영역을 추출한다. 셋째, 다중 스케일 폐-집중 패치를 사용하여 폐렴의 특징을 학습한다. 마지막으로 분류 성능을 향상시키기 위해 어텐션 모듈을 추가한 앙상블 모델을 사용한다. 실험 결과, CLAHE를 적용한 큰 크기의 패치 사용 시 정확도 92%로 원 영상 대비 5%p 향상된 성능을 보였다. 또한 큰 크기와 중간 크기의 패치를 앙상블한 제안 방법이 정확도 93%로 가장 좋은 성능을 보였다.
이중섭(C. S. Yi),심규진(K. J. Shim),이용훈(Y. H. Lee),정한식(H. S. Chung),정효민(H. M. Jeong) 한국동력기계공학회 2006 한국동력기계공학회 학술대회 논문집 Vol.- No.-
This research is focused on the application for the catalytic converter in the motorcycle engine. Present research model type is monolithic catalytic converter and this type has been widely used for satisfaction and the regulations of pollutant emissions in automobiles. The flow characteristics in a single monolith automotive catalytic converter were investigated by using a computational simulation method. The numerical mode with a general cartesian coordinates system is assumed as the steady and unsteady state, incompressible flow and κ - ε standard turbulence model. The inlet boundary condition of catalytic converter is assumed as a uniform distributed flow and it was varied from 2m/s to 20m/s. Also, boundary condition of unsteady state was a pressure condition by the experimental result. There has no chemical reaction in this study, so we discuss only to the effect of flow uniformity in the catalyst converter. Simulation results lead us to the conclusion that the fluid uniformity of megaphone type catalytic converter is higher than that of a base type one. By comparing with various types, megaphone type has more suitable for applicable to motorcycle.