• The Journal of Advanced Prosthodontics
• /
• 제15권2호
• /
• pp.72-79
• /
• 2023

PURPOSE. The purpose of this in vitro study was to compare the accuracy of various 3D printers and a milling machine. MATERIALS AND METHODS. The die model was designed using CAD (Autodesk Inventor 2018 sp3). The 30 ㎛ cement space was given to the die and the ideal crown of the mandibular left first molar was designed using CAD (ExoCAD). The crowns were produced using the milling machine (Imes-icore 250i) and the 3D printers (Zenith U, Zenith D, W11) and they were divided into four groups. In all groups, the interior of each crown was scanned (Identica blue) and superimposed (Geomagic Control X) with the previously designed die. The difference between the die and the actual crown was measured at specific points. The Kruskal-Wallis test, the Mann-Whitney test, and Bonferroni's method were performed with a statistical analysis software (P < .008 in inter-group comparison P < .001 in intra-group comparison). RESULTS. In all groups, the center of the occlusal area and the anti-rotational dimple area showed significantly greater difference and the marginal area showed the smallest difference comparatively. The mean value of the difference in each area and the sum of the differences were higher in order of W11, Imes-icore 250i, Zenith D, and Zenith U. CONCLUSION. The digital light processing (DLP) method shows higher accuracy compared to the sereolithography (SLA) method using the same resin material.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권5호
• /
• pp.1433-1449
• /
• 2023

In recent years, the rapid development of social networks has led to a rapid increase in the amount of information available on the Internet, which contains a large amount of sensitive information related to pornography, politics, and terrorism. In the aspect of sensitive image detection, the existing machine learning algorithms are confronted with problems such as large model size, long training time, and slow detection speed when auditing and supervising. In order to detect sensitive images more accurately and quickly, this paper proposes a multiclassification sensitive image detection method based on lightweight Convolutional Neural Network. On the basis of the EfficientNet model, this method combines the Ghost Module idea of the GhostNet model and adds the SE channel attention mechanism in the Ghost Module for feature extraction training. The experimental results on the sensitive image data set constructed in this paper show that the accuracy of the proposed method in sensitive information detection is 94.46% higher than that of the similar methods. Then, the model is pruned through an ablation experiment, and the activation function is replaced by Hard-Swish, which reduces the parameters of the original model by 54.67%. Under the condition of ensuring accuracy, the detection time of a single image is reduced from 8.88ms to 6.37ms. The results of the experiment demonstrate that the method put forward has successfully enhanced the precision of identifying multi-class sensitive images, significantly decreased the number of parameters in the model, and achieved higher accuracy than comparable algorithms while using a more lightweight model design.

• Restorative Dentistry and Endodontics
• /
• 제21권1호
• /
• pp.289-299
• /
• 1996

It is necessary to measure the length of a root canal in order to attain a satisfactory prognosis after root canal therapy. There are several methods for determining root canal length, such as tactile sensation by the dental practitioner, the utilization of x-ray film, and electronic root canal measurement. Among these, the electrical measurement methods, in which the impedence between the oral mucous membrane and periodontal membrane is determined, have advantages of simplicity and accuracy. During root canal treatment, the root canal contains a solution of high electrical conductivity such as pus, blood, sodium hypochlorite and so on. Recently a new electronic root canal measurement device of frequency-dependent type has been developed, which is capable of measuring the length of root canal under moist conditions. Endex and Root ZX, which are frequency-dependent type, were evaluated for accuracy of measuring root canal length in vivo by stereomicroscope. The result were as follows ; 1. 82.5% of Endex and 87.5% of Root ZX measured in the range of ${\pm}0.5$ mm from the apical foramen and both showed 57.5 % in the range of 0.1 mm to 0.5 mm. 2. Endex showed significantly higher accuracy in vital teeth than nonvital teeth(p<0.05). But in case of Root ZX, there was no significant difference between vital and nonvital teeth. 3. As a result of this study, there was no significant difference in accuracy between Endex and Root ZX, and both devices showed file passes the apical foramen in more than half of the cases, and it is thought that this must be considered clinically.

• 한국해안해양공학회지
• /
• 제15권3호
• /
• pp.186-189
• /
• 2003

방파제 피복석의 중량 산정을 위하여 여러 가지 경험식들이 개발되어져 왔다 Hudson 식은 많은 결함이 있음에도 불구하고, 형태가 간결하기 때문에 가장 널리 사용되고 있다. van der Meer 식은 가끔 크로스체크 용으로 사용되고 있지만, 식의 복잡성과 관련 변수들의 불확실성 때문에 아주 널리 사용되지는 않고 있다 한편 최근에 유(2003)는 'action slope'라는 새로운 무차원수를 사용하여 형태가 간결한 식을(이후 Yoo 식이라고 부름) 제안하였다. 본 연구에서는, 해안공학 기술자들에게 이 공식들의 사용에 대한 판단 기준을 제공하기 위하여, van der Meer가 1987년과 1988년에 보고한 실험 자료와 비교하여 이들의 정밀도를 평가하였다. 검토 결과, van der Meer 식이 가장 높은 정밀도를 보였으며, Hudson 식은 정밀도가 매우 낮아서 사용에 신중을 기해야 하는 것으로 나타났다. 한편, van der Meer 식보다 정밀도가 약간 낮기는 하지만 형태가 간결한 Yoo 식은 안전율을 약간 높게 잡으면 실무에 사용해도 무방한 것으로 판단된다.

• 한국측량학회지
• /
• 제40권1호
• /
• pp.15-22
• /
• 2022

본 연구에서는 딥러닝 객체탐지 기법의 정확도 향상을 위해 항공사진과 드론 영상을 대상으로 확대율 조건과 계절요인이 탐지정확도에 미치는 영향을 실험을 통해 분석하였다. 딥러닝 객체탐지기법 중 빠른 학습 속도와 높은 정확도를 나타내는 Mask R-CNN을 사용하여 탐지대상인 자동차를 픽셀 단위로 탐지하고자 하였다. '서울시 항공사진서비스'를 통해 화면 확대 레벨을 달리하며 학습 영상을 캡처하고 각각을 학습하여 정확도를 분석하였다. 실험결과에 따르면 확대 레벨이 높아질수록 mAP 평균이 60%, 67%, 75%로 높아졌다. 데이터 세트의 train, test 데이터의 확대율을 엇갈려서 배치한 경우에는 확대율이 매우 낮은 경우를 제외하고 저배율의 데이터를 train 데이터로, 고배율의 데이터를 test 데이터로 배치하였을 때 높은 mAP로 반대의 경우보다 20% 이상 차이를 보였다. 그리고 4개월의 시차로 계절적 차이를 두고 촬영한 드론 영상의 경우, 같은 시기 영상자료 학습결과가 평균 93%로 높은 정확도를 나타내어 계절적 차이도 학습에 영향을 주는 것을 확인되었다.

• 한국컴퓨터정보학회논문지
• /
• 제22권4호
• /
• pp.41-49
• /
• 2017

Word sense disambiguation(WSD) that determines the exact meaning of homonym which can be used in different meanings even in one form is very important to understand the semantical meaning of text document. Many recent researches on WSD have widely used NNLM(Neural Network Language Model) in which neural network is used to represent a document into vectors and to analyze its semantics. Among the previous WSD researches using NNLM, RNN(Recurrent Neural Network) model has better performance than other models because RNN model can reflect the occurrence order of words in addition to the word appearance information in a document. However, since RNN model uses only the forward order of word occurrences in a document, it is not able to reflect natural language's characteristics that later words can affect the meanings of the preceding words. In this paper, we propose a WSD scheme using Bidirectional RNN that can reflect not only the forward order but also the backward order of word occurrences in a document. From the experiments, the accuracy of the proposed model is higher than that of previous method using RNN. Hence, it is confirmed that bidirectional order information of word occurrences is useful for WSD in Korean language.

• 대한조선학회논문집
• /
• 제42권2호
• /
• pp.113-123
• /
• 2005

A higher order panel method based on B-spline representation for both the geometry and the solution is developed for the analysis of steady flow around marine propellers. The self-influence functions due to the normal dipole and the source are desingularized through the quadratic transformation, and then shown to be evaluated using conventional numerical quadrature. By selecting a proper order for numerical quadrature, the accuracy of the present method can be increased to the machine limit. The far- and near-field influences are shown to be evaluated based on the same far-field approximation, but the near-field solution requires subdividing the panels into smaller subpanels continuously, which can be effectively implemented due to the B-spline representation of the geometry. A null pressure jump Kutta condition at the trailing edge is found to be effective in stabilizing the solution process and in predicting the correct solution. Numerical experiments indicate that the present method is robust and predicts the pressure distribution on the blade surface, including very close to the tip and trailing edge regions, with far fewer panels than existing low order panel methods.

• 한국전산유체공학회지
• /
• 제9권1호
• /
• pp.34-40
• /
• 2004

Recently with growth in the capability of super computers and Parallel computers, shape design optimization is becoming easible for real problems. Also, Computational Fluid Dynamics(CFD) techniques have been improved for higher reliability and higher accuracy. In the shape design optimization, analysis solvers and optimization schemes are essential. In this work, the Roe's 2nd-order Upwind TVD scheme and DADI time march with multigrid were used for the flow solution with the Euler equation and FDM(Finite Differenciation Method), GA(Genetic Algorithm) and Kriging were used for the design optimization. Kriging were applied to 2-D airfoil design optimization and compared with FDM and GA's results. When Kriging is applied to the nonlinear problems, satisfactory results were obtained. From the result design optimization by Kriging method appeared as good as other methods.

• 제어로봇시스템학회논문지
• /
• 제10권9호
• /
• pp.840-846
• /
• 2004

In this study, the performance of a heavy load pointing system has been investigated. The PI controller are being widely used in industrial application because of simple, cheap, and excellent performance. However, the requirement for control precision becomes higher and higher, as well as the plants becomes more and more complex. In order to achieve the satisfied control performance, we have to consider the affection of nonlinear factor contained in plant. In this paper, the neural-PI control law have been evaluated. The proposed controller is compared with the existing controllers through simulations, and the results show that the pointing accuracy of the proposed control system is improved against the disturbance induced by vehicle running on the bump course.

• Steel and Composite Structures
• /
• 제25권4호
• /
• pp.389-401
• /
• 2017

In this work, a new higher shear deformation theory (HSDT) is developed for the free vibration and buckling of functionally graded (FG) sandwich plates. The proposed theory presents a new displacement field by using undetermined integral terms. Only four unknowns are employed in this theory, which is less than the classical first shear deformation theory (FSDT) and others HSDTs. Equations of motion are obtained via Hamilton's principle. The analytical solutions of FG sandwich plates are determined by employing the Navier method. A good agreement between the computed results and the available solutions of existing HSDTs is found to prove the accuracy of the developed theory.