• Journal of the Korean Institute of Intelligent Systems
• /
• v.12 no.6
• /
• pp.511-516
• /
• 2002

The neural networks may have problem such that the amount of calculation for the network learning goes too big according to the dimension of the dimension. To overcome this problem, the wavelet neural networks(WNN) which use the orthogonal basis function in the hidden node are proposed. One can compose wavelet functions as activation functions in the WNN by determining the scale and center of wavelet function. In this paper, when we compose the WNN using wavelet functions, we set a single scale function as a node function together. We intend that one scale function approximates the target function roughly, the other wavelet functions approximate it finely During the determination of the parameters, the wavelet functions can be determined by the global search for solutions suitable for the suggested problem using the genetic algorithm and finally, we use the back-propagation algorithm in the learning of the weights.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.11
• /
• pp.747-756
• /
• 2013

It is important to reduce unnecessary overhead in sensor network using battery. In addition encryption is important because of necessity of security. Since unavoidable overhead occurs in case of encryption, security and overhead are in trade-off condition. In this paper, we use a concept called trust to reduce the encryption overhead. We reduce overhead by controlling encryption key sizes while maintaining the security level where high and low trust nodes are mixed. We simulated and compared normal encryption and trust value based encryption. As a result, the latter has lower execution time and overhead. If we define a standard of trust levels considering purpose and circumstances of real network, we can use constrained resources efficiently in sensor network.

• Journal of KIISE:Databases
• /
• v.34 no.6
• /
• pp.473-482
• /
• 2007

Network-based IDS(Intrusion Detection System) gathers network packet data and analyzes them into attack or normal. They raise alarm when possible intrusion happens. But they often output a large amount of low-level of incomplete alert information. Consequently, a large amount of incomplete alert information that can be unmanageable and also be mixed with false alerts can prevent intrusion response systems and security administrator from adequately understanding and analyzing the state of network security, and initiating appropriate response in a timely fashion. So it is important for the security administrator to reduce the redundancy of alerts, integrate and correlate security alerts, construct attack scenarios and present high-level aggregated information. False alarm rate is the ratio between the number of normal connections that are incorrectly misclassified as attacks and the total number of normal connections. In this paper we propose a false alarm classification model to reduce the false alarm rate using classification analysis of data mining techniques. The proposed model can classify the alarms from the intrusion detection systems into false alert or true attack. Our approach is useful to reduce false alerts and to improve the detection rate of network-based intrusion detection systems.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.47 no.7
• /
• pp.51-58
• /
• 2010

Wireless Mesh Network has drawn much attention due to wide area service coverage with low system cost. However, there is a bottleneck problem in wireless mesh network since the traffic is aggregated into a gateway. Placement of multi-radio can easy the bottleneck problem, but without careful design it results in unnecessary system cost increasement. In this paper, we propose a system cost minimization through differential antenna placement where optimum antenna placement is determined by the required wireless link capacity. With CPLEX program, optimum number of antennae is determined as a function of local user traffic and gateway capacity. From numerical analysis, it is confirmed that our proposed model can solve bottleneck problem, and at the same time save the system cost.

• Journal of the Korea Society of Computer and Information
• /
• v.19 no.1
• /
• pp.85-94
• /
• 2014

This paper studied the detection technique using file DNA-based behavior pattern analysis in order to minimize damage to user system by malicious programs before signature or security patch is released. The file DNA-based detection technique was applied to defend against zero day attack and to minimize false detection, by remedying weaknesses of the conventional network-based packet detection technique and process-based detection technique. For the file DNA-based detection technique, abnormal behaviors of malware were splitted into network-related behaviors and process-related behaviors. This technique was employed to check and block crucial behaviors of process and network behaviors operating in user system, according to the fixed conditions, to analyze the similarity of behavior patterns of malware, based on the file DNA which process behaviors and network behaviors are mixed, and to deal with it rapidly through hazard warning and cut-off.

• Journal of KIISE:Computing Practices and Letters
• /
• v.14 no.4
• /
• pp.352-362
• /
• 2008

In the mobile ad hoc network, network nodes are continuously capable of movement, but have battery limitations and these characteristics of MANET cause network link failure. Therefore, when performing ubiquitous services with the application of MANET, a technique is required to improve the packet transmission rates as to enable continuous packet transmission while flexibly adapting to topology changes caused by the mobility of nodes. In this paper, we are proposing PAOMDV(Possibility based AOMDV) and OLSR-ME(OLSR with Modified Energy Efficiency) which are mixed-type protocols to solve the problem for design a more efficient protocol. Especially by classifying mobile ad hoc networks as clusters, OLSR-ME protocol is implemented on communication within the clusters, while PAOMDV is used in protocols between clusters. In addition, a selection algorithm for forwarding nodes that responding packet relays in the case of increased hops was proposed for designing a more efficient protocol than existing protocols. We verified the performance of the proposed protocols through a comparative experiment with AODV, AOMDV, ZRP, and the existing protocols. We confirmed the results of the experiment revealed that the proposed protocol had the best packet transmission rate, the shortest delay between nodes, and showed better results in routing overhead.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.10a
• /
• pp.1137-1140
• /
• 2019

가상현실(Virtual Reality, VR), 증강현실(Augmented Reality, AR), 혼합현실(Mixed Reality, MR) 분야에 유용한 인간 컴퓨터 인터페이스 기술은 필수적이다. 특히 휴먼 손동작 인식 기술은 직관적인 상호작용을 가능하게 하여, 다양한 분야에서 편리한 컨트롤러로 사용할 수 있다. 본 연구에서는 뎁스 영상 기반의 1 인칭 시점 손동작 인식을 위하여 손동작 데이터베이스 생성 시스템을 구축하여, 손동작 인식기 학습에 필요한 1 인칭(Egocentric View Point) 데이터베이스를 촬영하여 제작한다. 그리고 모바일 Head Mounted Device(HMD) VR 을 위한 뎁스 영상 기반 1 인칭 시점 손동작 인식(Hand Pose Recognition, HPR) 딥러닝 Deep Residual Network 를 구현한다. 최종적으로, 안드로이드 모바일 디바이스에 학습된 Residual Network Regressor 를 이식하고 모바일 VR 에 실시간 손동작 인식 시스템을 구동하여, 모바일 VR 상 실시간 3D 손동작 인식을 가상 물체와의 상호작용을 통하여 확인 한다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.3
• /
• pp.427-432
• /
• 2021

In this paper, we conduct performance analysis in automatic modulation classification of unknown communication signal to identify its modulation types based on deep neural network. The modulation classification performance was verified using time domain digital sample data of the modulated signal, frequency domain data to which FFT was applied, and time and frequency domain mixed data as neural network input data. For 11 types of analog and digitally modulated signals, the modulation classification performance was verified in various SNR environments ranging from -20 to 18 dB and reason for false classification was analyzed. In addition, by checking the learning speed according to the type of input data for neural network, proposed method is effective for constructing an practical automatic modulation recognition system that require a lot of time to learn.

• Current Optics and Photonics
• /
• v.7 no.3
• /
• pp.237-247
• /
• 2023

In middle- and long-distance imaging systems, due to the atmospheric turbulence caused by temperature, wind speed, humidity, and so on, light waves propagating in the air are distorted, resulting in image-quality degradation such as geometric deformation and fuzziness. In remote sensing, astronomical observation, and traffic monitoring, image information loss due to degradation causes huge losses, so effective restoration of degraded images is very important. To restore images degraded by atmospheric turbulence, an image-restoration method based on improved compound multibranch feature fusion (CMFNetPro) was proposed. Based on the CMFNet network, an efficient channel-attention mechanism was used to replace the channel-attention mechanism to improve image quality and network efficiency. In the experiment, two-dimensional random distortion vector fields were used to construct two turbulent datasets with different degrees of distortion, based on the Google Landmarks Dataset v2 dataset. The experimental results showed that compared to the CMFNet, DeblurGAN-v2, and MIMO-UNet models, the proposed CMFNetPro network achieves better performance in both quality and training cost of turbulent-image restoration. In the mixed training, CMFNetPro was 1.2391 dB (weak turbulence), 0.8602 dB (strong turbulence) respectively higher in terms of peak signal-to-noise ratio and 0.0015 (weak turbulence), 0.0136 (strong turbulence) respectively higher in terms of structure similarity compared to CMFNet. CMFNetPro was 14.4 hours faster compared to the CMFNet. This provides a feasible scheme for turbulent-image restoration based on deep learning.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.25 no.1B
• /
• pp.15-30
• /
• 2000

The important considerations in the design of fiber-optic networks are reliability and survivability preparing against a failure. The SDH(Synchronous Digital Hierarchy), the international standard of optical transmission, offers several network reconfiguration methods that enable network to be automatically restored from failure. One of the methods is the SHR(Self Healing Ring), which is a ring topology system. Most network providers have constructed their backbone networks with SHR architecture since it can provide survivability economically. The network architecture has eventually evolved into a multi-ring network comprised of interconnected rings. This paper addresses multi-ring network design problems is to minimize ring-construction cost. This problem can be formulated with MIP(mixed integer programming) model. However, it is difficult to solve the model within reasonable computing time on a large scale network because the model is NP-complete. Furthermore, in practice we should consider the problem of routing demands on rings to minimize total cost. This routing problem involves multiplex bundling at the intermediate nodes. A family of heuristic algorithms is presented for this problem. These algorithms include gateway selection and routing of inter-ring demands as well as load balancing on single rings. The developed heuristic algorithms are applied to some network provider's regional and long-distance transmission networks. We show an example of ring design and compare it with another ring topology design. Finally, we analysis the effect bundling.