• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.60 no.2
• /
• pp.398-406
• /
• 2011

In this paper, we introduce a new architecture of PSO-based Polynomial Neural Networks (PNN) and discuss its comprehensive design methodology. The conventional PNN is based on a extended Group Method of Data Handling (GMDH) method, and utilized the polynomial order (viz. linear, quadratic, and modified quadratic) as well as the number of node inputs fixed (selected in advance by designer) at Polynomial Neurons located in each layer through a growth process of the network. Moreover it does not guarantee that the conventional PNN generated through learning results in the optimal network architecture. The PSO-based PNN results in a structurally optimized structure and comes with a higher level of flexibility that the one encountered in the conventional PNN. The PSO-based design procedure being applied at each layer of PNN leads to the selection of preferred PNs with specific local characteristics (such as the number of input variables, input variables, and the order of the polynomial) available within the PNN. In the sequel, two general optimization mechanisms of the PSO-based PNN are explored: the structural optimization is realized via PSO whereas in case of the parametric optimization we proceed with a standard least square method-based learning. To evaluate the performance of the PSO-based PNN, the model is experimented with using Gas furnace process data, and pH neutralization process data. For the characteristic analysis of the given entire data with non-linearity and the construction of efficient model, the given entire system data is partitioned into two type such as Division I(Training dataset and Testing dataset) and Division II(Training dataset, Validation dataset, and Testing dataset). A comparative analysis shows that the proposed PSO-based PNN is model with higher accuracy as well as more superb predictive capability than other intelligent models presented previously.

• Journal of Information Processing Systems
• /
• v.16 no.1
• /
• pp.6-29
• /
• 2020

Biometrics identification using multiple modalities has attracted the attention of many researchers as it produces more robust and trustworthy results than single modality biometrics. In this paper, we present a novel multimodal recognition system that trains a deep learning network to automatically learn features after extracting multiple biometric modalities from a single data source, i.e., facial video clips. Utilizing different modalities, i.e., left ear, left profile face, frontal face, right profile face, and right ear, present in the facial video clips, we train supervised denoising auto-encoders to automatically extract robust and non-redundant features. The automatically learned features are then used to train modality specific sparse classifiers to perform the multimodal recognition. Moreover, the proposed technique has proven robust when some of the above modalities were missing during the testing. The proposed system has three main components that are responsible for detection, which consists of modality specific detectors to automatically detect images of different modalities present in facial video clips; feature selection, which uses supervised denoising sparse auto-encoders network to capture discriminative representations that are robust to the illumination and pose variations; and classification, which consists of a set of modality specific sparse representation classifiers for unimodal recognition, followed by score level fusion of the recognition results of the available modalities. Experiments conducted on the constrained facial video dataset (WVU) and the unconstrained facial video dataset (HONDA/UCSD), resulted in a 99.17% and 97.14% Rank-1 recognition rates, respectively. The multimodal recognition accuracy demonstrates the superiority and robustness of the proposed approach irrespective of the illumination, non-planar movement, and pose variations present in the video clips even in the situation of missing modalities.

• Journal of Satellite, Information and Communications
• /
• v.6 no.2
• /
• pp.72-77
• /
• 2011

Spectrum sensing is a critical functionality of CR network; it allow secondary user to detect spectral holes and to opportunistically use under-utilized frequency bands without causing harmful interference to primary use. Recently, wideband service has been increase for processing abundance of data traffic. So CR network needs a realizable implementation design of spectrum sensing for wideband. To get high resolution performance of wideband sensing must precede algorithm processing for reliability signal detection. By the way, the performance of spectrum sensing can be degraded due to fading and shadowing. In order to overcome this problem, we propose system model of wideband sensing scheme on energy detected collaborative technique. we divide wideband into narrowbands and use narrowbands to detect signal excepting some narrowbands including bad channel through the CSI. And we simulate and analyze in terms of detection probability with various SNR.

• The KIPS Transactions:PartC
• /
• v.17C no.2
• /
• pp.189-196
• /
• 2010

The LEACH does not use flooding method for data transmission and this makes low power consumption. So performance of the WSN is increased. On the other hand, QoS based algorithm which use restricted flooding method in WSN also achieves low power consuming rate by reducing the number of nodes that are participated in routing path selection. But when the data is delivered to the sink node, the LEACH choose a routing path which has a small hop count. And it leads that the performance of the entire network is worse. In the paper we propose a QoS based energy efficient clustering and routing algorithm in WSN. I classify the type of packet with two classes, based on the energy efficiency that is the most important issue in WSN. We provide the differentiated services according to the different type of packet. Simulation results evaluated by the NS-2 show that proposed algorithm extended the network lifetime 2.47 times at average. And each of the case in the class 1 and class 2 data packet, the throughput is improved 312% and 61% each.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.12 no.12
• /
• pp.2259-2264
• /
• 2008

Significant genes are defined as genes in which the expression level characterizes a specific experimental condition. Such genes in which the expression levels differ significantly between different groups are highly informative relevant to the studied phenomenon. In this paper, first the system can detect informative genes by similarity scale combination method being proposed in this paper after normalizing data with methods that are the most widely used among several normalization methods proposed the while. And it compare and analyze a performance of each of normalization methods with multi-perceptron neural network layer. The Result classifying in Multi-Perceptron neural network classifier for selected 200 genes using combination of PC(Pearson correlation coefficient) and ED(Euclidean distance coefficient) after Lowess normalization represented the improved classification performance of 98.84%.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.4
• /
• pp.47-53
• /
• 2020

Blockchain is a distributed shared ledger that transparently manages information through verification and agreement between nodes connected to a distributed network. Recently, cases of data management among authorized agencies based on private blockchain are increasing. In this paper, we investigated the application cases and technical processes of PBFT, the representative consensus algorithm of private blockchain, and proposed a modified PBFT algorithm that enables efficient consensus by simplifying duplicate verification and consensus processes that occur during PBFT processing. The algorithm proposed in this paper goes through the process of selecting a delegation node through an authoritative node and can increase the safety of the delegation node selection process by considering an efficient re-election algorithm for candidate nodes. By utilizing this research, it is possible to reduce the burden on the network communication cost of the consensus process and effectively process the final consensus process between nodes.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.1
• /
• pp.19-26
• /
• 2019

In this paper, the classification performance of learning algorithms is compared for TAB digit recognition. The TAB digits that are segmented from TAB musical notes contain TAB lines and musical symbols. The labeling method and non-linear filter are designed and applied to extract fret digits only. The shift operation of the 4 directions is applied to generate more data. The selected models are Bayesian classifier, support vector machine, prototype based learning, multi-layer perceptron, and convolutional neural network. The result shows that the mean accuracy of the Bayesian classifier is about 85.0% while that of the others reaches more than 99.0%. In addition, the convolutional neural network outperforms the others in terms of generalization and the step of the data preprocessing.

• Journal of Advanced Navigation Technology
• /
• v.22 no.6
• /
• pp.577-584
• /
• 2018

GNSS based transport infrastructure cluster is to broadcast satellite navigation correction information and integrity information capable of precise positioning for land transport users. This makes it possible to do lane-level positioning reliably. However, in order to provide the lane-level positioning and correction information service nationwide, new station sites selection and to build GNSS stations have a heavy cost and a burden for a considerable period of time. In this paper, we propose the cluster design criteria and national-wide network-based precise positioning for land transportation (NETPPI-LT) cluster design for a cluster-based precise positioning. Furthermore, it is analyzed the precise positioning pre-performance of this cluster design based on the spatial error and verified its suitability as the precise positioning pre-performance of the cluster design.

• Journal of Digital Convergence
• /
• v.16 no.11
• /
• pp.185-199
• /
• 2018

The purpose of this study is to derive the core technologies of VR using patent analysis and to explore the direction of social and public interest in VR using news analysis. In Study 1, we derived keywords using the frequency of words in patent texts, and we compared by company, year, and technical classification. Netminer, a network analysis program, was used to analyze the IPC codes of patents. In Study 2, we analyzed news articles using T-LAB program. TF-IDF was used as a keyword selection method and chi-square and association index algorithms were used to extract the words most relevant to VR. Through this study, we confirmed that VR is a fusion technology including optics, head mounted display (HMD), data analysis, electric and electronic technology, and found that optical technology is the central technology among the technologies currently being developed. In addition, through news articles, we found that the society and the public are interested in the formation and growth of VR suppliers and markets, and VR should be developed on the basis of user experience.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.15 no.2
• /
• pp.53-64
• /
• 2005

Previous secure routing protocols for wireless sensor networks assume that a sink is static. In many cases, however, a sink operated by man or vehicle is moving. A mobile sink creates a lot of technical problems such as reconfiguration of routing path exposure of sink location. and selection of secure access point node, which are not considered by many previous researches. In this paper, we propose a new secure routing scheme for solving such problems using hi-directional hash chain and delegation nodes of grid structure. This scheme provides a secure routing path and prevents attacker from recognizing the location of a mobile sink in sensor networks. This new method reduces the resource requirements compared to the cashed routing schemes. Simulation results also show that the system is secure and efficient enough.