• Journal of the Korean Data and Information Science Society
• /
• v.27 no.2
• /
• pp.327-335
• /
• 2016

The People's Republic of China has vigorously been pursuing the internationalization of the Chinese Yuan or Renminbi after the financial crisis of 2008. In this view, an abrupt increase of use of the Chinese Yuan in the onshore and offshore markets are important milestones to be one of important currencies. One of the most frequently used methods to forecast volatility is GARCH model. Since a prediction error of the GARCH model has been reported quite high, a lot of efforts have been made to improve forecasting capability of the GARCH model. In this paper, we have proposed MLP-GARCH and a DL-GARCH by employing Artificial Neural Network to the GARCH. In an application to forecasting Chinese Yuan volatility, we have successfully shown their overall outperformance in forecasting over the GARCH.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.2
• /
• pp.800-816
• /
• 2018

In recent years, video surveillance research has been able to recognize various behaviors of pedestrians and analyze the overall situation of objects by combining image analysis technology and deep learning method. Human Activity Recognition (HAR), which is important issue in video surveillance research, is a field to detect abnormal behavior of pedestrians in CCTV environment. In order to recognize human behavior, it is necessary to detect the human in the image and to estimate the pose from the detected human. In this paper, we propose a novel approach for 2D Human Pose Estimation based on object detection using RGB-D information. By adding depth information to the RGB information that has some limitation in detecting object due to lack of topological information, we can improve the detecting accuracy. Subsequently, the rescaled region of the detected object is applied to ConVol.utional Pose Machines (CPM) which is a sequential prediction structure based on ConVol.utional Neural Network. We utilize CPM to generate belief maps to predict the positions of keypoint representing human body parts and to estimate human pose by detecting 14 key body points. From the experimental results, we can prove that the proposed method detects target objects robustly in occlusion. It is also possible to perform 2D human pose estimation by providing an accurately detected region as an input of the CPM. As for the future work, we will estimate the 3D human pose by mapping the 2D coordinate information on the body part onto the 3D space. Consequently, we can provide useful human behavior information in the research of HAR.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.4
• /
• pp.1721-1737
• /
• 2020

Human apparent age estimation has become a popular research topic and attracted great attention in recent years due to its wide applications, such as personal security and law enforcement. To achieve the goal of age estimation, a large number of methods have been pro-posed, where the models derived through the cumulative attribute coding achieve promised performance by preserving the neighbor-similarity of ages. However, these methods afore-mentioned ignore the geometric structure of extracted facial features. Indeed, the geometric structure of data greatly affects the accuracy of prediction. To this end, we propose an age estimation algorithm through joint feature selection and manifold learning paradigms, so-called Feature-selected and Geometry-preserved Least Square Regression (FGLSR). Based on this, our proposed method, compared with the others, not only preserves the geometry structures within facial representations, but also selects the discriminative features. Moreover, a deep learning extension based FGLSR is proposed later, namely Feature selected and Geometry preserved Neural Network (FGNN). Finally, related experiments are conducted on Morph2 and FG-Net datasets for FGLSR and on Morph2 datasets for FGNN. Experimental results testify our method achieve the best performances.

• Journal of Broadcast Engineering
• /
• v.25 no.2
• /
• pp.242-251
• /
• 2020

Recently, many deep convolutional neural networks for image super-resolution have been studied. Existing deep learning-based super-resolution algorithms are architecture that up-samples the resolution at the end of the network. The post-upsampling architecture has an inefficient structure at large scaling factor result of predicting a lot of information for mapping from low-resolution to high-resolution at once. In this paper, we propose a single image super-resolution using Channel Attention Residual Dense Block based on an iterative up-down sampling architecture. The proposed algorithm efficiently predicts the mapping relationship between low-resolution and high-resolution, and shows up to 0.14dB performance improvement and enhanced subjective image quality compared to the existing algorithm at large scaling factor result.

• Food Science of Animal Resources
• /
• v.38 no.2
• /
• pp.362-375
• /
• 2018

This study proposed a rapid microscopic examination method for pork freshness evaluation by using the self-assembled hyperspectral microscopic imaging (HMI) system with the help of feature extraction algorithm and pattern recognition methods. Pork samples were stored for different days ranging from 0 to 5 days and the freshness of samples was divided into three levels which were determined by total volatile basic nitrogen (TVB-N) content. Meanwhile, hyperspectral microscopic images of samples were acquired by HMI system and processed by the following steps for the further analysis. Firstly, characteristic hyperspectral microscopic images were extracted by using principal component analysis (PCA) and then texture features were selected based on the gray level co-occurrence matrix (GLCM). Next, features data were reduced dimensionality by fisher discriminant analysis (FDA) for further building classification model. Finally, compared with linear discriminant analysis (LDA) model and support vector machine (SVM) model, good back propagation artificial neural network (BP-ANN) model obtained the best freshness classification with a 100 % accuracy rating based on the extracted data. The results confirm that the fabricated HMI system combined with multivariate algorithms has ability to evaluate the fresh degree of pork accurately in the microscopic level, which plays an important role in animal food quality control.

• Journal of Korean Society for Quality Management
• /
• v.46 no.3
• /
• pp.739-752
• /
• 2018

Purpose: The purpose of this study was to investigate the relationship between kiln processing parameters and NOx emissions that occur in the sintering and calcination steps of the cement manufacturing process and to derive the main factors responsible for producing emissions outside emission limit criteria, as determined by category models and classification rules, using data mining techniques. The results from this study are expected to be useful as guidelines for NOx emission control standards. Methods: Data were collected from Precalciner Kiln No.3 used in one of the domestic cement plants in Korea. Thirty-four independent variables affecting NOx generation and dependent variables that exceeded or were below the NOx emiision limit (>1 and <0, respectively) were examined during kiln processing. These data were used to construct a detection model of NOx emission, in which emissions exceeded or were below the set limits. The model was validated using SPSS MODELER 18.0, artificial neural network, decision treee (C5.0), and logistic regression analysis data mining techniques. Results: The decision tree (C5.0) algorithm best represented NOx emission behavior and was used to identify 10 processing variables that resulted in NOx emissions outside limit criteria. Conclusion: The results of this study indicate that the decision tree (C5.0) can be applied for real-time monitoring and management of NOx emissions during the cement firing process to satisfy NOx emission control standards and to provide for a more eco-friendly cement product.

• Journal of Welding and Joining
• /
• v.34 no.1
• /
• pp.54-58
• /
• 2016

In this research paper, suggest method of generate same bead as an actual measurement data in virtual welding conditions, exploit morphology information of the bead that acquired through robot welding. It has many multiple risk factors to Beginners welding training, by we make possible to train welding in virtual reality, we can reduce welding training risk and welding material to exploit bead visualization algorithm that we suggest so it will be expected to achieve educational, environmental and economical effect. The proposed method is acquire data to each case performing robot welding by set the voltage, current, working angle, process angle, speed and arc length of welding condition value. As Welding condition value is most important thing in decide bead form, we would selected one of baseline each item and then acquired metal followed another factors change. Welding type is FCAW, SMAW and TIG. When welding trainee perform the training, it's difficult to save all of changed information into database likewise working angle, process angle, speed and arc length. So not saving data into database are applying the method to infer the form of bead using a neural network algorithm. The way of bead's visualization is applying the spline algorithm. To accurately represent Morphological information of the bead, requires much of morphological information, so it can occur problem to save into database that is why we using the spline algorithm. By applying the spline algorithm, it can make simplified data and generate accurate bead shape. Through the research paper, the shape of bead generated by the virtual reality was able to improve the accuracy when compared using the form of bead generated by the robot welding to using the morphological information of the bead generated through the robot welding. By express the accurate shape of bead and so can reduce the difference of the actual welding training and virtual welding, it was confirmed that it can be performed safety and high effective virtual welding education.

• Journal of Institute of Control, Robotics and Systems
• /
• v.6 no.2
• /
• pp.207-216
• /
• 2000

This paper introduces an effective system design method to develop a customer oriented product using a design optimization process and to select a set of critical design paramenters,. The process results in the development of a successful product satisfying customer needs and reducing development risk. The proposed scheme adopted a five step QFD(Quality Function Deployment) in order to extract design parameters from customer needs and evaluated their priority using risk factors for extracted design parameters. In this process we determine critical design parameters and allocate them to subsystem designers. Subsequently design engineers develop and test the product based on these parameters. These design parameters capture the characteristics of customer needs in terms of performance cost and schedule in the process of QFD, The subsequent risk management task ensures the minimum risk approach in the presence of design parameter uncertainty. An application of this approach was demonstrated in the development of weld quality monitoring system. Dominant design parameters affect linearity characteristics of weld defect feature vectors. Therefore it simplifies the algorithm for adopting pattern classification of feature vectors and improves the accuracy of recognition rate of weld defect and the real time response of the defect detection in the performance. Additionally the development cost decreases by using DSP board for low speed because of reducing CPU's load adopting algorithm in classifying weld defects. It also reduces the cost by using the single sensor to measure weld defects. Furthermore the synergy effect derived from the critical design parameters improves the detection rate of weld defects by 15% when compared with the implementation using the non-critical design parameters. It also result in 30% saving in development cost./ The overall results are close to 95% customer level showing the effectiveness of the proposed development approach.

• 한국HCI학회:학술대회논문집
• /
• 2006.02a
• /
• pp.519-525
• /
• 2006

Advances in information and telecommunication technology increasingly reveal the potential of computer supported education. However, most computer supported learning systems until recently did not pay much attention to different characteristics of individual learners. Intelligent learning environments adaptive to learner's preferences and tasks are desired. Each learner has different preferences and needs, so it is very crucial to provide the different styles of learners with different learning environments that are more preferred and more efficient to them. This paper reports a study of the intelligent learning environment where the learner's preferences are diagnosed using learner models, and then user interfaces are customized in an adaptive manner to accommodate the preferences. In this research, the learning user interfaces were designed based on a learning-style model by Felder & Silverman, so that different learner preferences are revealed through user interactions with the system. Then, a learning style modeling is done from learner behavior patterns using Decision Tree and Neural Network approaches. In this way, an intelligent learning system adaptive to learning styles can be built. Further research efforts are being made to accommodate various other kinds of learner characteristics such as emotion and motivation as well as learning mastery in providing adaptive learning support.

• Journal of Korea Multimedia Society
• /
• v.13 no.4
• /
• pp.504-511
• /
• 2010

In this paper, we propose the face region recognition system for City-Security CCTV(Closed Circuit Television) using hippocampal neural network which is modelling of human brain's hippocampus. This system is composed of feature extraction, learning and recognition part. The feature extraction part is constructed using PCA(Principal Component Analysis) and LDA(Linear Discriminants Analysis). In the learning part, it can label the features of the image-data which are inputted according to the order of hippocampal neuron structure to reaction-pattern according to the adjustment of a good impression in a dentate gyrus and remove the noise through the auto-associative memory in the CA3 region. In the CA1 region receiving the information of the CA3, it can make long-term memory learned by neuron. Experiments confirm the each recognition rate, that are shape change and light change. The experimental results show that we can compare a feature extraction and learning method proposed in this paper of any other methods, and we can confirm that the proposed method is superior to existing methods.