• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권3호
• /
• pp.1188-1203
• /
• 2020

The human expression contains a lot of information that can be used to detect complex conditions such as pain and fatigue. After deep learning became the mainstream method, the traditional feature extraction method no longer has advantages. However, in order to achieve higher accuracy, researchers continue to stack the number of layers of the neural network, which makes the real-time performance of the model weak. Therefore, this paper proposed an expression recognition framework based on densely concatenated convolutional neural networks to balance accuracy and latency and apply it to humanoid robots. The techniques of feature reuse and parameter compression in the framework improved the learning ability of the model and greatly reduced the parameters. Experiments showed that the proposed model can reduce tens of times the parameters at the expense of little accuracy.

• Structural Engineering and Mechanics
• /
• 제73권1호
• /
• pp.1-16
• /
• 2020

Finite elements based on the partition of unity (PU) approximation have powerful capabilities for p-adaptivity and solutions with high smoothness without remeshing of the domain. Recently, the PU approximation was successfully applied to the three-node shell finite element, properly eliminating transverse shear locking and showing excellent convergence properties and solution accuracy. However, the enrichment with the PU approximation results in a significant increase in the number of degrees of freedom; therefore, it requires greater computational cost, thus making it less suitable for practical engineering. To circumvent this disadvantage, we propose a new strategy to decrease the total number of degrees of freedom in the existing PU-based shell element, without loss of optimal convergence and accuracy. To alleviate the locking phenomenon, we use the method of mixed interpolation of tensorial components and perform convergence studies to show the accuracy and capability of the proposed shell element. The excellent performances of the new shell elements are illustrated in three benchmark problems.

• 디지털산업정보학회논문지
• /
• 제15권3호
• /
• pp.197-206
• /
• 2019

Telemarketing has been used in finance due to the reduction of offline channels. In order to select telemarketing target customers, various machine learning techniques have emerged to maximize the effect of minimum cost. However, there are problems that the class imbalance, which the number of marketing success customers is smaller than the number of failed customers, and the recall rate is lower than accuracy. In this paper, we propose a method that solve the imbalanced class problem and increase the recall rate to improve the efficiency. The hybrid sampling method is applied to balance the data in the class, and the stacked deep network is applied to improve the recall and precision as well as the accuracy. The proposed method is applied to actual bank telemarketing data. As a result of the comparison experiment, the accuracy, the recall, and the precision is improved higher than that of the conventional methods.

• 한국정밀공학회지
• /
• 제24권6호
• /
• pp.45-50
• /
• 2007

The aspheric lens has become the most popular optical component used in various optical devices such as digital cameras, pick-up lenses, printers, copiers etc. Using aspheric lenses not only miniaturizes and reduces the weight of products, but also lower prices and higher field angles can be realized. Additionally, plastic lenses are being changed to glass lenses more recently because of low accuracy, low acid-resistance and low thermal-resistance in the plastic lenses. Currently, one fabrication method of glass lenses is using a glass-mold method with a high precision mold core for mass production. In this paper, DOE (Design Of Experiments) and compensation machining were adopted to improve the surface roughness and the form accuracy of the mold core. The DOE has been done in order to discover the optimal grinding conditions which minimize the surface roughness with factors such as work spindle revolution, turbine spindle revolution, federate and cutting depth. And the compensation machining is used to generate high form accuracy of the mold core. From various experiments and analyses, we could obtain the best surface roughness 5 nm in Ra, form accuracy $0.167\;{\mu}m$ in PV.

• 소성∙가공
• /
• 제13권5호
• /
• pp.429-434
• /
• 2004

Manufacturing process for milli components has recently gained researcher's focus with the increasing tendency toward highly integrated and micro-scaled parts for electronic devices. The milli-components need more precise manufacturing process than the conventional manufacturing process since the parts require higher dimensional accuracy than the conventional ones. In order to enhance the forming accuracy and productivity, various forming procedures proposed and studied by many researchers. In this paper, forming analysis of milli-components has been studied with a new micro-former. In modeling of progressive dies, multi-stage forming sequence has been analyzed with finite element analysis by LS-DYNA3D. The analysis proposes the sequential die and part shapes with the corresponding punch force and dimensional accuracy. The analysis also considers the effect of elastic dies on the dimensional accuracy of the formed parts. The analysis result demonstrates that the elastic analysis in the milli-forming process is indispensable for accurate forming analysis. The analysis procedure in the paper will provide good information in design of a new micro-former and milli-component

• 한국환경보건학회지
• /
• 제22권2호
• /
• pp.32-42
• /
• 1996

This study was performed to evaluate accuracy and precision of filter method and impinger method for analyzin airborne isocyanates in mixture (2, 6-TDI, HDI, 2, 4-TDI, MDI). Filter method was performed using the OSHA Method 42 and impinger method using the NIOSH Method 5521. The samples were analyzed by high performance liquid chromatography-ultraviolet detector (HPLC-UVD). After the optimum operating conditions for each method were investigated, samples with various concentration levels were quantified at the conditions. The precision was expressed by the pooled coefficient of variation(C.V.) and the accuracy by overall accuracy. The results are summarized as follows: 1. The optimum condition of filter method was determined at 35/65 ACN/buffer (0.01 M ammonium acetate) in mobile phase. And in case of impinger method, it was at 30/70 ACN/buffer(0.2 M sodium acetate). The effect of concentrations of acetate on the separation of the peaks was not significant, but, the effect of ACN/buffer ratio was significant. 2. The correlation coefficients for the two methods were above 0.9 in all isocyanate compounds. Average recovery efficiencies for 2, 6-TDI, HDI, 2, 4-TDI and MDI in filter method were 92.4%, 102.6%, 87.3% and 101.0%, respectively. Those in impinger method were 106.6%, 106.7%, 99.0% and 103.6%, respectively. As a result, the recovery efficiency of impinger method was higher than those of filter method in analyzing isocyanate compounds. 3. The pooled coefficients of variations of the methods were slightly higher than expected. The overall accuracies of the methods were within $\pm 25%$ for each isocyanate compound. Since these results satisfy NIOSH criteria, the accuracy of the experiment is appropriate. 4. As seen above, impinger method is more efficient than filter method. But, there are many disadvantages in impinger method. Therefore, solid sorbent such as a glass fiber filter must be developed in order to have the high efficiency not less than that of impinger method in the future.

• 한국항공우주학회지
• /
• 제35권11호
• /
• pp.1043-1051
• /
• 2007

디지털 태양센서는 CMOS 이미지 센서에 맺힌 태양광 이미지를 이용하여 태양광의 입사 각도를 계산한다. 이를 위해서는 태양광 이미지의 정확한 중심점을 찾아야하며 따라서 정밀한 중심점 추정은 디지털 태양센서 개발에서 가장 중요한 요소가 된다. 중심점을 찾기 위해서 가장 일반적으로 쓰이는 중심 알고리즘은 thresholding 방법이며 가장 단순하고 구현하기 쉽다. 또 다른 알고리즘으로는 이미지 처리를 이용하는 image filtering 방법이 있다. 하지만 이러한 방법들은 태양센서 정밀도가 이미지 센서에서 획득한 태양광 강도(intensity) 데이터의 노이즈에 영향을 많이 받으며, 특히 thresholding 방법의 경우 threshold 값에 따라 정밀도가 바뀌기 때문에 효과적인 threshold 값을 정하기 어려운 단점이 있다. 따라서 본 논문에서는 태양광 이미지의 중심점을 구하기 위해서 태양광 이미지 모델을 이용하는 template 방법을 제시하고 성능을 비교 분석하였다. 제안한 template 방법은 thereshold, image filtering 방법과 달리 비교적 높은 정밀도를 가지며, 특히 노이즈 수준에 관계없이 거의 일정한 수준의 정밀도를 가지는 장점이 있어 신뢰성이 높다.

• International Journal of Computer Science & Network Security
• /
• 제22권10호
• /
• pp.246-256
• /
• 2022

Recently, parents and teachers consider physical education as a minor subject for students in elementary and secondary schools. Physical education performance has become increasingly significant as parents and schools pay more attention to physical schooling. The sports mining with distribution analysis model considers different factors, including the games, comments, conversations, and connection made on numerous sports interests. Using different machine learning/deep learning approach, children's athletic and academic interests can be tracked over the course of their academic lives. There have been a number of studies that have focused on predicting the success of students in higher education. Sports interest prediction research at the secondary level is uncommon, but the secondary level is often used as a benchmark to describe students' educational development at higher levels. An Automated Student Interest Prediction on Sports Mining using DL Based Bi-directional Long Short-Term Memory model (BiLSTM) is presented in this article. Pre-processing of data, interest classification, and parameter tweaking are all the essential operations of the proposed model. Initially, data augmentation is used to expand the dataset's size. Secondly, a BiLSTM model is used to predict and classify user interests. Adagrad optimizer is employed for hyperparameter optimization. In order to test the model's performance, a dataset is used and the results are analysed using precision, recall, accuracy and F-measure. The proposed model achieved 95% accuracy on 400^th instances, where the existing techniques achieved 93.20% accuracy for the same. The proposed model achieved 95% of accuracy and precision for 60%-40% data, where the existing models achieved 93% for accuracy and precision.

• 대한기계학회논문집A
• /
• 제31권1호
• /
• pp.121-128
• /
• 2007

Enhanced lower order shear deformation theory is developed in this study. Generally, lower order theories are not adequate to predict accurate deformation and stress distribution through the thickness of laminated plate. For the accurate prediction of detailed stress and deformation distributions through the thickness, higher order zigzag theories have been proposed. However, in most cases, simplified zigzag higher order theory requires $C_1$, shape functions in finite element implementation. In commercial FE softwares, $C_1$, shape functions are not so common in plate and shell analysis. Thus zigzag theories are useful for the highly accurate prediction of thick composite behaviors but they are not practical in the sense that they cannot be used conveniently in the commercial package. In practice, iso-parametric $C_0$ plate model is the standard model for the analysis and design of composite laminated plates and shells. Thus in the present study, an enhanced lower order shear deformation theory is developed. The proposed theory requires only $C_0$ shape function in FE implementation. The least-squared energy error between the lower order theory and higher order theory is minimized. An enhanced lower order shear deformation theory(ELSDT) in this paper is proposed for smart structure under complex loadings. The ELSDT is constructed by the strain energy transformation and fully coupled mechanical, electric loading cases are studied. In order to obtain accurate prediction, zigzag in-plane displacement and transverse normal deformation are considered in the deformation Held. In the electric behavior, open-circuit condition as well as closed-circuit condition is considered. Through the numerous examples, the accuracy and robustness of present theory are demonstrated.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2005년도 Proceedings of ISRS 2005
• /
• pp.116-119
• /
• 2005

As to the synthetical estimation of land covering parameters or the compounded land covering classification for multi-resolution satellite data, former researches mainly adopted linear or nonlinear regression models to describe the regression relationship of land covering parameters caused by the degradation of spatial resolution, in order to improve the retrieval accuracy of global land covering parameters based on 1;he lower resolution satellite data. However, these methods can't authentically represent the complementary characteristics of spatial resolutions among different satellite data at arithmetic level. To resolve the problem above, a new compounded land covering classification method at arithmetic level for multi-resolution satellite data is proposed in this .paper. Firstly, on the basis of unsupervised clustering analysis of the higher resolution satellite data, the likelihood distribution scatterplot of each cover type is obtained according to multiple-to-single spatial correspondence between the higher and lower resolution satellite data in some local test regions, then Parzen window approach is adopted to derive the real likelihood functions from the scatterplots, and finally the likelihood functions are extended from the local test regions to the full covering area of the lower resolution satellite data and the global covering area of the lower resolution satellite is classified under the maximum likelihood rule. Some experimental results indicate that this proposed compounded method can improve the classification accuracy of large-scale lower resolution satellite data with the support of some local-area higher resolution satellite data.