• Journal of the Korea Society of Computer and Information
• /
• v.26 no.3
• /
• pp.51-57
• /
• 2021

In this paper, we propose a method to find the DoF(Depth of field) that is blurred in an image by focusing and out-focusing the camera through a efficient convolutional neural network. Our approach uses the RGB channel-based cross-correlation filter to efficiently classify the DoF region from the image and build data for learning in the convolutional neural network. A data pair of the training data is established between the image and the DoF weighted map. Data used for learning uses DoF weight maps extracted by cross-correlation filters, and uses the result of applying the smoothing process to increase the convergence rate in the network learning stage. The DoF weighted image obtained as the test result stably finds the DoF region in the input image. As a result, the proposed method can be used in various places such as NPR(Non-photorealistic rendering) rendering and object detection by using the DoF area as the user's ROI(Region of interest).

• Journal of the Korean Geographical Society
• /
• v.45 no.6
• /
• pp.788-805
• /
• 2010

The present USN (Ubiquitous Sensor Networks) node deployment practices have many limitations in terms of positional connectivity. The aim of this research was to minimize a redundancy of USN route nodes, by integrating spatially weighted parameters such as visibility, proximity to cell center, road density, building density and cell overlapping ratio into a comprehensive GIS database. This spatially weighted approach made it possible to reduce the number of route nodes (11) required in the study site as compared to that of the grid network method (24). The field test for RSSI (Received Signal Strength Indicator) indicates that the spatially weighted deployment could comply with the quality assurance standard for node connectivity, and that reduced route nodes do not show a significant degree of signal fluctuation for different site conditions. This study demonstrated that the spatially weighted deployment can be used to minimize a redundancy of USN route nodes in a routine manner, and the quantitative evidence removing a redundancy of USN route nodes could be utilized as major tools to ensure the strong signal in the USN, that is frequently encountered in real applications.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.18 no.10
• /
• pp.1422-1432
• /
• 1993

Nonlinear mapping function of the HCNN( Hidden Control Neural Network ) can change over time to model the temporal variability of a speech signal by combining the nonlinear prediction of conventional neural networks with the segmentation capability of HMM. We have two things in this paper. first, we showed that the performance of the HCNN is better than that of HMM. Second, the HCNN with its prediction error measure given by weighted distance is proposed to use suitable distance measure for the HCNN, and then we showed that the superiority of the proposed system for speaker-independent speech recognition tasks. Weighted distance considers the differences between the variances of each component of the feature vector extraced from the speech data. Speaker-independent Korean digit recognition experiment showed that the recognition rate of 95%was obtained for the HCNN with Euclidean distance. This result is 1.28% higher than HMM, and shows that the HCNN which models the dynamical system is superior to HMM which is based on the statistical restrictions. And we obtained 97.35% for the HCNN with weighted distance, which is 2.35% better than the HCNN with Euclidean distance. The reason why the HCNN with weighted distance shows better performance is as follows : it reduces the variations of the recognition error rate over different speakers by increasing the recognition rate for the speakers who have many misclassified utterances. So we can conclude that the HCNN with weighted distance is more suit-able for speaker-independent speech recognition tasks.

• Journal of radiological science and technology
• /
• v.26 no.1
• /
• pp.71-82
• /
• 2003

In this paper, I proposed a classifier of liver cirrhotic step using T1-weighted MRI(magnetic resonance imaging) and hierarchical neural network. The data sets for classification of each stage, which were normal, 1type, 2type and 3type, were obtained in Pusan National University Hospital from June 2001 to december 2001. And the number of data was 46. We extracted liver region and nodule region from T1-weighted MR liver image. Then objective interpretation classifier of liver cirrhotic steps in T1-weighted MR liver images. Liver cirrhosis classifier implemented using hierarchical neural network which gray-level analysis and texture feature descriptors to distinguish normal liver and 3 types of liver cirrhosis. Then proposed Neural network classifier teamed through error back-propagation algorithm. A classifying result shows that recognition rate of normal is 100%, 1type is 82.3%, 2type is 86.7%, 3type is 83.7%. The recognition ratio very high, when compared between the result of obtained quantified data to that of doctors decision data and neural network classifier value. If enough data is offered and other parameter is considered, this paper according to we expected that neural network as well as human experts and could be useful as clinical decision support tool for liver cirrhosis patients.

• Journal of the Korean Society for Library and Information Science
• /
• v.49 no.3
• /
• pp.457-488
• /
• 2015

In this study, we analyzed the current research state of Library and Information science in top 20 journals from 1990 to 2015, in subject and method perspectives. We developed weighted subject-method network to investigate on centralities of a subject and a method as well as their relations. This network is composed of subject nodes and method nodes and gives a weight on each node by topic occurrence. As a result, for 25 years, management information system, information need analysis, bibliometrics, information policy were top topics. Modeling, literature review, scientific research impact analysis, web data analysis were top methods. A recent rise of text mining is highlighted. We also analyzed communities made from the past 25 years and the recent 5 years. Bibliometrics is extending its field by applying various network analyzing algorithms. Text mining is specialized in medical information system and user interface. This result identifies the interests of excellent studies in Library and Information Science. It also can be fundamental resource for the development of Library and Information Science.

• Journal of Intelligence and Information Systems
• /
• v.21 no.3
• /
• pp.37-52
• /
• 2015

The recent global trends display expansion and growing solidity in both cooperative collaboration between industry, education, and research and R&D network systems. A greater support for the network and cooperative research sector would open greater possibilities for the evolution of new scholar and industrial fields and the development of new theories evoked from synergized educational research. Similarly, the national need for a strategy that can most efficiently and effectively support R&D network that are established through the government's R&D project research is on the rise. Despite the growing urgency, due to the habitual dependency on simple individual personal information data regarding R&D industry participants and generalized statistical data references, the policies concerning network system are disappointing and inadequate. Accordingly, analyses of the relationships involved for each subject who is participating in the R&D industry was conducted and on the foundation of an educational-industrial-research network system, possible changes within and of the network that may arise were predicted. To predict the R&D network transitions, Common Neighbor and Jaccard's Coefficient models were designated as the basic foundational models, upon which a new prediction model was proposed to address the limitations of the two aforementioned former models and to increase the accuracy of Link Prediction, with which a comparative analysis was made between the two models. Through the effective predictions regarding R&D network changes and transitions, such study result serves as a stepping-stone for an establishment of a prospective strategy that supports a desirable educational-industrial-research network and proposes a measure to promote the national policy to one that can effectively and efficiently sponsor integrated R&D industries. Though both weighted applications of Common Neighbor and Jaccard's Coefficient models provided positive outcomes, improved accuracy was comparatively more prevalent in the weighted Common Neighbor. An un-weighted Common Neighbor model predicted 650 out of 4,136 whereas a weighted Common Neighbor model predicted 50 more results at a total of 700 predictions. While the Jaccard's model demonstrated slight performance improvements in numeric terms, the differences were found to be insignificant.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.6B
• /
• pp.358-366
• /
• 2007

Recently, FCC(Federal Communications Commission) has considered for that unlicensed device leases licensed devices' channel to overcome shortage of communication channels. Therefore, IEEE 802.22 WRAN(Wireless Regional Area Networks) working group progresses CR (Cognitive Radio) technique that is able to sense and adopt void channels that are not being occupied by the licensed devices. Channel selection is of the utmost importance because it can affect the whole system performance in CR network. Thus, we propose a policy-based dynamic channel selection architecture for cognitive radio network to achieve an efficient communication. We propose three kinds of method for channel selection; the first one is weighted channel selection, the second one is sequential channel selection, and the last one is combined channel selection. We can obtain the optimum channel list and allocates channels dynamically using the proposed protocol.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.38 no.6
• /
• pp.677-682
• /
• 2014

In the present study, a decision tree and artificial neural network were used to determine critical design parameters for lithium ion batteries and compare their performances. First, a design method that used a decision tree-artificial neural network model was used to determine the major design factors among early pole plate design factors that showed nonlinearity. Then, the artificial neural network was used to implement a weighted value analysis of the importance of the design factors and their effect on the current density. The second method involved the use of an artificial neural network model to construct artificial networks without separate determinations of the major early design factors to analyze the connections and weighted values related to the current density.

• Journal of KIISE:Software and Applications
• /
• v.33 no.8
• /
• pp.692-702
• /
• 2006

In this study, a weighted fuzzy min-max (WFMM) neural network model for pattern classification is proposed. The model has a modified structure of FMM neural network in which the weight concept is added to represent the frequency factor of feature values in a learning data set. First we present in this paper a new activation function of the network which is defined as a hyperbox membership function. Then we introduce a new learning algorithm for the model that consists of three kinds of processes: hyperbox creation/expansion, hyperbox overlap test, and hyperbox contraction. A weight adaptation rule considering the frequency factors is defined for the learning process. Finally we describe a feature analysis technique using the proposed model. Four kinds of relevance factors among feature values, feature types, hyperboxes and patterns classes are proposed to analyze relative importance of each feature in a given problem. Two types of practical applications, Fisher's Iris data and Cleveland medical data, have been used for the experiments. Through the experimental results, the effectiveness of the proposed method is discussed.

• Journal of information and communication convergence engineering
• /
• v.16 no.1
• /
• pp.6-11
• /
• 2018

It is important to find the random estimation points in wireless sensor network. A link quality indicator (LQI) is part of a network management service that is suitable for a ZigBee network and can be used for localization. The current quality of the received signal is referred as LQI. It is a technique to demodulate the received signal by accumulating the magnitude of the error between ideal constellations and the received signal. This proposed model accepts any number of random estimation point in the network and calculated its nearest anchor centroid node pair. Coordinates of the LQI sphere are calculated from the pair and are added iteratively to the initially estimated point. With the help of the LQI and weighted centroid localization, the proposed system finds the position of target node more accurately than the existing system by solving the problems related to higher error in terms of the distance and the deployment of nodes.