• 대한치과교정학회지
• /
• 제52권1호
• /
• pp.3-19
• /
• 2022

Objective: The purpose of this study was to investigate the accuracy of one-step automated orthodontic diagnosis of skeletodental discrepancies using a convolutional neural network (CNN) and lateral cephalogram images with different qualities from nationwide multi-hospitals. Methods: Among 2,174 lateral cephalograms, 1,993 cephalograms from two hospitals were used for training and internal test sets and 181 cephalograms from eight other hospitals were used for an external test set. They were divided into three classification groups according to anteroposterior skeletal discrepancies (Class I, II, and III), vertical skeletal discrepancies (normodivergent, hypodivergent, and hyperdivergent patterns), and vertical dental discrepancies (normal overbite, deep bite, and open bite) as a gold standard. Pre-trained DenseNet-169 was used as a CNN classifier model. Diagnostic performance was evaluated by receiver operating characteristic (ROC) analysis, t-stochastic neighbor embedding (t-SNE), and gradient-weighted class activation mapping (Grad-CAM). Results: In the ROC analysis, the mean area under the curve and the mean accuracy of all classifications were high with both internal and external test sets (all, > 0.89 and > 0.80). In the t-SNE analysis, our model succeeded in creating good separation between three classification groups. Grad-CAM figures showed differences in the location and size of the focus areas between three classification groups in each diagnosis. Conclusions: Since the accuracy of our model was validated with both internal and external test sets, it shows the possible usefulness of a one-step automated orthodontic diagnosis tool using a CNN model. However, it still needs technical improvement in terms of classifying vertical dental discrepancies.

• 한국전산응용수학회:학술대회논문집
• /
• 한국전산응용수학회 2003년도 KSCAM 학술발표회 프로그램 및 초록집
• /
• pp.1-1
• /
• 2003

In this talk, a reduced-order modeling methodology based on centroidal Voronoi tessellations (CVT's)is introduced. CVT's are special Voronoi tessellations for which the generators of the Voronoi diagram are also the centers of mass (means) of the corresponding Voronoi cells. The discrete data sets, CVT's are closely related to the h-means clustering techniques. Even with the use of good mesh generators, discretization schemes, and solution algorithms, the computational simulation of complex, turbulent, or chaotic systems still remains a formidable endeavor. For example, typical finite element codes may require many thousands of degrees of freedom for the accurate simulation of fluid flows. The situation is even worse for optimization problems for which multiple solutions of the complex state system are usually required or in feedback control problems for which real-time solutions of the complex state system are needed. There hava been many studies devoted to the development, testing, and use of reduced-order models for complex systems such as unsteady fluid flows. The types of reduced-ordered models that we study are those attempt to determine accurate approximate solutions of a complex system using very few degrees of freedom. To do so, such models have to use basis functions that are in some way intimately connected to the problem being approximated. Once a very low-dimensional reduced basis has been determined, one can employ it to solve the complex system by applying, e.g., a Galerkin method. In general, reduced bases are globally supported so that the discrete systems are dense; however, if the reduced basis is of very low dimension, one does not care about the lack of sparsity in the discrete system. A discussion of reduced-ordering modeling for complex systems such as fluid flows is given to provide a context for the application of reduced-order bases. Then, detailed descriptions of CVT-based reduced-order bases and how they can be constructed of complex systems are given. Subsequently, some concrete incompressible flow examples are used to illustrate the construction and use of CVT-based reduced-order bases. The CVT-based reduced-order modeling methodology is shown to be effective for these examples and is also shown to be inexpensive to apply compared to other reduced-order methods.

• 정보과학회 논문지
• /
• 제44권1호
• /
• pp.71-77
• /
• 2017

스마트폰과 같은 GPS센서가 장착된 모바일 기기가 널리 보급되고 있다. 이러한 추세에 따라 트위터나 페이스북과 같은 소셜 미디어에서 GEO태그가 된 메시지 (즉, GPS 위치를 갖는 멀티미디어 메시지)를 손쉽게 작성할 수 있게 되었고 게시자의 위치정보를 포함하는 공간데이터가 급증하고 있다. 그러나 이러한 공간데이터에서 항상 위치 정보와 게시글 내용이 명시적인 연관성을 갖고 있지는 않다. 때문에 키워드와 위치정보 분포의 관련성에 따라 검색결과를 재구성할 필요가 있다. 우리는 크기가 가장 작은 k개의 점 데이터를 포함하는 최소경계사각형(MBR)을 찾음으로써 데이터가 가장 밀집된 사각형을 찾으며, 이는 위치 검색시스템에서 유용하게 사용될 수 있다. 본 논문은 최소경계사각형과 같이 거리가 가까운 2차원 공간데이터의 묶음을 찾기 위해 효율적인 알고리즘을 제안하였고 합성데이터와 실제 데이터를 이용한 실험을 통해 알고리즘의 효율성을 검증하였다.

• 건축역사연구
• /
• 제11권4호
• /
• pp.99-115
• /
• 2002

Centered on Durand and Le Corbusier, this study analyses the changing status of the grid and axis in modern architecture. In the Renaissance, the taxis grid operated as a contour grid, defining the elements and space of the building as part of closed harmonized world. In his Pre'cis des lec., ons d'architecture, Durand provides the most explicit demonstration of a new modem grid in which its lines function as spatial and structural axes. In principle these axes are coordinates for the placements of a priori elements but in Beaux-Arts practice, as Durand himself acknowledged, they involve a simultaneous process in which the spatial axis sets up the basic parti and the structural axis is developed into the building's poche'. As a coordinate, Durand's grid provides a place for the 'subject' to enter the architectural process. At the same time, it is the object of the subject's gaze, the dense site of the subject's transformative actions. Though Le Corbusier is noted for his frequent attacks on the academic system, his architecture should be seen within the continuity of the classical tradition. He redefines the Beaux-Arts axis as a moving and seeing observer, and continues the discipline of the plan, the essential discipline of the Beaux-Arts system. In his dialectics, an intellectual scheme which extends to his commentators, the intention and will of the subject must come in tune with the objective material form of the building. Like Durand, Le Corbusier's axis provides the medium for the subject to enter. Unlike the Beaux-Arts system, however, Le Corbusier's mobile subject no longer has a holistic view of the building previously provided by the central axis. If there is a parti for Le Corbusier, it consists of the domino grid as a potential, but nonetheless, tangible form. In comparison with the Beaux-Arts structural grid, his gaze no longer lingers on their lines because they no longer constitute a formal process tied to the development of a thick articulated structure. Le Corbusier's grid constitutes a 'loose' form, one that breaks down the hierarchical nature of the Beaux-Arts system.

• Asian Pacific Journal of Cancer Prevention
• /
• 제14권7호
• /
• pp.4141-4149
• /
• 2013

Background: Mammography is considered the gold standard of breast cancer mass screening and many countries have implemented this as an established breast cancer screening strategy. However, although the incidence of breast cancer and racial characteristics are different between Western and Asian countries, many Asian countries adopted mammography for mass screening. Therefore, the objective of this research was to determine whether mammography mass screening is cost-effective for both Western and Asian countries. Materials and Methods: A systematic review was performed of 17 national mammography cost-effectiveness data sets. Per capita gross domestic product (GDP), breast cancer incidence rate, and the most optimal cost-effectiveness results [cost per life year saved (LYS)] of a mammography screening strategy for each data set were extracted. The CE/per capita GDP ratio is used to compare the cost-effectiveness of mammography by countries. Non-parametric regression was used to find a cut-off point which indicated the breast cancer incidence rate boundary line determining whether mammography screening is cost-effective or not. Results: We found that the cost-effective cut-off point of breast cancer incidence rate was 45.04; it exactly divided countries into Western and Asian countries (p<0.0014). Conclusions: Mammography screening is cost-effective in most of Western countries, but not in Asian countries. The reason for this result may be the issues of incidence rate or racial characteristics, such as dense breast tissue. The results indicate that mammography screening should be adopted prudently in Asian countries and other countries with low incidence rates.

• 한국수자원학회논문집
• /
• 제48권3호
• /
• pp.233-243
• /
• 2015

이군집 응집현상은 수자원환경에서 점착성 유사가 결합-해체의 과정을 통해 응집핵-응집체의 이군집 입자크기분포 (Biomodal Floc Size Distribution)를 형성하는 일련의 과정을 의미한다. 본 연구는 저난류 및 고난류 두 가지 조건에서 수행한 응집-침전관 실험결과를 바탕으로 이군집 응집모형(TCPBE: Two Class Population Balance Equation)의 적용성을 단일군집 응집모형(SCPBE: Single Class Population Balance Equation) 및 다군집 응집모형(MCPBE: Multi Class Population Balance Equation)과 비교 평가하였다. 기존 SCPBE에 비하여, TCPBE는 응집핵-응집체의 상호작용 및 침강속도차에 따른 응집 기작을 모의할 수 있었다. 또한, 3개의 연립미분방정식을 가진 TCPBE는 30개 미분방정식을 가진 다군집 응집모형(MCPBE: Multi Class Population Balance Equation)과 대등한 모의 결과를 나타내었다. 따라서 TCPBE는 이군집 응집현상을 모의 할 수 있는 가장 단순한 모델로 검증되었고, 향후 수자원환경이나 수처리 공정에 다양하게 적용할 수 있으리라 판단된다.

• 한국통신학회논문지
• /
• 제26권11A호
• /
• pp.1872-1879
• /
• 2001

본 논문에서는 다시점 영상에서 영상의 특성을 고려하여 효율적으로 미세 변이를 추정하고 중간 시점 영상을 합성하는 알고리듬을 제안하며 이를 모의 실험을 통해 검증한다. 제안한 방법은 영역분할 양방향 화소정합을 통해 변이 추정의 수행 속도를 향상시키는 동시에 신뢰도를 높이며, 적절한 비용함수의 제안과 유사 영역에 대한 정합 창 확장 알고리듬, 변이 평활화와 불확실 영역 변이 할당 알고리듬을 통해 잘못 할당된 변이와 불확실 영역을 제거함으로서 변이의 신뢰도를 더욱 높일 수 있다. 이렇게 추정된 변이는 다시점 영상의 중간 시점 영상 합성을 위해 사용된다. 제안된 방식을 통해 기존의 방식들보다 더욱 안정적인 변이 정보를 얻을 수 있었고, 합성된 중간 시점 영상도 객관적으로나 주관적으로나 더욱 좋은 결과를 보였으며, 또한 전체적인 수행 시간도 줄어들어 더 효율적인 알고리듬임을 확인할 수 있었다.

• 한국세라믹학회지
• /
• 제51권6호
• /
• pp.598-604
• /
• 2014

Considerable research efforts have been explored attempting to enhance the thermal durability of thermal barrier coatings (TBCs) at the high operating temperatures of gas turbines. In this study, the suspension plasma spray (SPS) process was applied to produce TBCs with a segmented structure by using an yttria-stabilized zirconia (YSZ) suspension. Four different experiment sets were carried out by controlling the ratio between surface roughness of the bond coat and feed stock size ($R_a/D_{50}$) in order to examine the effect of $R_a/D_{50}$ ratio on the microstructure of SPS-prepared coatings. When the $R_a/D_{50}$ had a high value of 11.8, a deposited thick coating turned out to have a cone-type columnar microstructure. In contrast, at the low $R_a/D_{50}$ values of 2.9 and 0.18, a deposited thick coating appeared to have a dense, vertically-cracked microstructure. However, with the very low $R_a/D_{50}$ value of 0.05 the coating was delaminated.

• Advances in nano research
• /
• 제12권2호
• /
• pp.185-195
• /
• 2022

Deep learning is another field of artificial intelligence (AI) utilized for computer aided diagnosis (CAD) and image processing in scientific research. Considering numerous mechanical repetitive tasks, reading image slices need time and improper with geographical limits, so the counting of image information is hard due to its strong subjectivity that raise the error ratio in misdiagnosis. Regarding the highest mortality rate of Lung cancer, there is a need for biopsy for determining its class for additional treatment. Deep learning has recently given strong tools in diagnose of lung cancer and making therapeutic regimen. However, identifying the pathological lung cancer's class by CT images in beginning phase because of the absence of powerful AI models and public training data set is difficult. Convolutional Neural Network (CNN) was proposed with its essential function in recognizing the pathological CT images. 472 patients subjected to staging FDG-PET/CT were selected in 2 months prior to surgery or biopsy. CNN was developed and showed the accuracy of 87%, 69%, and 69% in training, validation, and test sets, respectively, for T1-T2 and T3-T4 lung cancer classification. Subsequently, CNN (or deep learning) could improve the CT images' data set, indicating that the application of classifiers is adequate to accomplish better exactness in distinguishing pathological CT images that performs better than few deep learning models, such as ResNet-34, Alex Net, and Dense Net with or without Soft max weights.

• 지질공학
• /
• 제27권4호
• /
• pp.475-487
• /
• 2017

본 연구에서는 산사태 모형토조 실험을 통해 강우침투에 의해 나타나는 산사태의 발생특성 및 지반재료의 변화특성에 관하여 분석하였다. 실험장치는 모형토조, 강우재현장치, 계측장치로 구성되어 있으며, 인공사면 상부에서 200 mm/hr의 극한강우를 살수함으로써 산사태를 유발하였다. 모관흡수력과 체적함수비 계측장치는 인공사면의 천부(GL-0.2 m), 중부(GL-0.4 m), 심부(GL-0.6 m)의 각 심도별로 3세트씩 설치하였으며, 실험은 화강암 풍화토, 편마암 풍화토, 이암 풍화토 각각에 대한 현장조건과 상대적으로 느슨한 조건 및 조밀한 조건으로 나누어 실험을 진행하였다. 분석결과, 극한강우에 의하여 실험 초반에는 사면 표층에서 세굴현상이 우세하게 나타나다가 이후 국지적으로 발생한 횡적인 인장균열면을 따라서 산사태가 발생했다. 산사태는 강우침투에 따른 습윤전선(wetting front)의 전이로 인해 천부에서부터 심부로 점차 확장되는 천층파괴(shallow failure)의 형태를 띰과 동시에 사면의 선단부(toe part)에서부터 정상부(crest part)로 점차 전이되는 후퇴성붕괴(retrogressive failure)의 양상을 보였다. 강우침투에 따른 포화영역에서 모관흡수력은 아무런 전조현상 없이 급격하게 감소하는 반면에, 체적함수비는 점진적으로 증가하다가 최대값에 도달하여 이 값을 일정시간 유지한 뒤 산사태 발생 시 급격히 감소하는 것으로 나타났다.