• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2019년도 춘계학술대회
• /
• pp.432-435
• /
• 2019

Convolutional Neural Networks (CNNs) have recently been gaining popularity in the medical image analysis field because of their image segmentation capabilities. In this paper, we present a CNN that performs automated brain tumor segmentations of sparsely annotated 3D Magnetic Resonance Imaging (MRI) scans. Our CNN is based on 3D U-net architecture, and it includes separate Dilated and Depth-wise Convolutions. It is fully-trained on the BraTS 2018 data set, and it produces more accurate results even when compared to the winners of the BraTS 2017 competition despite having a significantly smaller amount of parameters.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2009년도 춘계학술발표대회
• /
• pp.842-845
• /
• 2009

Wireless ad hoc networks often suffer from rapidly degrading performance with the number of user increases in the network. One of the major reasons for this rapid degradation of performance is the fact that users are sharing a single channel. Obviously, the problem of using single shared channel schemes is that the probability of collision increases with the number of nodes. Fortunately, it is possible to solve this problem with multi-channel approaches. Due to the especial properties of multiple channels, using the multiple channels is more efficient than single channel because it enhances the capacity of the channel and reduces the error rate during data transmission. Some multi-channel schemes us one dedicated channel for control packets and one separate channel for data transmissions. On the other hand, another protocols use more than two channels for data transmissions. This paper summarizes six multiple channel protocols based on these two kinds of schemes. Then we compare them and discuss the research challenge of multiple channel protocols.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권4호
• /
• pp.1817-1836
• /
• 2020

Network convergence is considered as one of the key solutions to the problem of achieving future high-capacity and reliable communications. This approach overcomes the limitations of separate wireless technologies. Efficient interface selection is one of the most important issues in convergence networks. This paper solves the problem faced by users of selecting the most appropriate interface in the heterogeneous radio-access network (RAN) environment. Our proposed scheme combines a hierarchical evaluation of networks and game theory to solve the network-selection problem. Instead, of considering a fixed weight system while ranking the networks, the proposed scheme considers the service requirements, as well as static and dynamic network attributes. The best network is selected for a particular service request. To establish a hierarchy among the network-evaluation criteria for service requests, an analytical hierarchy process (AHP) is used. To determine the optimum network selection, the network hierarchy is combined with game theory. AHP attains the network hierarchy. The weights of different access networks for a service are calculated. It is performed by combining AHP scores considering user's experienced static network attributes and dynamic radio parameters. This paper provides a strategic game. In this game, the network scores of service requests for various RANs and the user's willingness to pay for these services are used to model a network-versus-user game. The Nash equilibria signify those access networks that are chosen by individual user and result maximum payoff. The examples for the interface selection illustrate the effectiveness of the proposed scheme.

• 대한전기학회논문지:시스템및제어부문D
• /
• 제53권8호
• /
• pp.561-570
• /
• 2004

In this paper, new architectures and comprehensive design methodologies of Genetic Algorithms(GAs) based Genetically optimized Neurofuzzy Networks(GoNFN) are introduced, and a series of numeric experiments are carried out. The proposed GoNFN is based on the rule-based Neurofuzzy Networks(NFN) with the extended structure of the premise and the consequence parts of fuzzy rules being formed within the networks. The premise part of the fuzzy rules are designed by using space partitioning in terms of fuzzy sets defined in individual variables. In the consequence part of the fuzzy rules, three different forms of the regression polynomials such as constant, linear and quadratic are taken into consideration. The structure and parameters of the proposed GoNFN are optimized by GAs. GAs being a global optimization technique determines optimal parameters in a vast search space. But it cannot effectively avoid a large amount of time-consuming iteration because GAs finds optimal parameters by using a given space. To alleviate the problems, the dynamic search-based GAs is introduced to lead to rapidly optimal convergence over a limited region or a boundary condition. In a nutshell, the objective of this study is to develop a general design methodology o GAs-based GoNFN modeling, come up a logic-based structure of such model and propose a comprehensive evolutionary development environment in which the optimization of the model can be efficiently carried out both at the structural as well as parametric level for overall optimization by utilizing the separate or consecutive tuning technology. To evaluate the performance of the proposed GoNFN, the models are experimented with the use of several representative numerical examples.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2004년도 ICCAS
• /
• pp.1067-1071
• /
• 2004

In this paper, the performance enhancement algorithm of channel allocation for voice and data transmission in cellular networks is proposed. The voice activity detection has been applied to dynamic channel allocation procedure to detect and separate the silence and speech among conversation periods. Hence a data user can use the silent period of an active voice channel to transmit its information. To control the selecting of channel allocation policies, the information of number of data in transmission waiting queue has been determined in order to accept the performance measurement. In the simulation results, the improvement of the performance shows via the quality of services, which are an average delay in queue, a blocking probability, and an impact of the proposed scheme is presented in the system.

• Journal of Electrical Engineering and Technology
• /
• 제8권5호
• /
• pp.1012-1020
• /
• 2013

In this work, Self-adaptive Differential Evolutionary (SaDE) algorithm is proposed to solve Optimal Location and Size of Capacitor (OLSC) problem in radial distribution networks. To obtain the SaDE algorithm, two improvements have been applied on control parameters of mutation and crossover operators. To expand the study, three load conditions have been considered, i.e., constant, varying and effective loads. Objective function is introduced for the load conditions. The annual cost is fitness of problem, in addition to this cost, CPU time, voltage profile, active power loss and total installed capacitor banks and their related costs have been used for comparisons. To confirm the ability of each improvements of SaDE, the improvements are studied both in separate and simultaneous conditions. To verify the effectiveness of the proposed algorithm, it is tested on IEEE 10-bus and 34-bus radial distribution networks and compared with other approaches.

• 한국통신학회논문지
• /
• 제21권8호
• /
• pp.1928-1936
• /
• 1996

In this paper, we propose an adaptive bandwidth allocation algorithm for the transmission of VBR video through ATM Networks. In order to evaluate the required bandwidth for the VBR video, the characteristics of the compressed VBR video generated by the two-layered coder are analyzed with variations in the number of GOP(N), quantizer scale(q), and the number of low-frequency DCT coefficients(.betha.). The two-layer coder which is used to separate from the number of DCT coefficients is designed to transmit the VBR video efficiently. The compressed data generated by the two-layer coder are splitted into the high priority and low priority cells. If congestion is occurred in ATM networks, the minimum image quality is maintained by the high priority cells. The required bandwidth for VBR video is estimated with a prediction algorithm using the scene anframe correlations as well as the statistical properties of the VBR video sources. Strong correlation among the adjacent slices in a frame represents by the scene correlation andstrong correlation among the frames is represented by the frame correlation. The performance of the bandwidth allocation scheme proposed is evaluated in terms of the bandwith utilization, cell loss rate, and SNR with variations in q, n, .betha.. Simulation rewsults shown that the proposed scheme is superior to the conventional methods.

• Journal of Communications and Networks
• /
• 제13권2호
• /
• pp.132-141
• /
• 2011

IEEE 802.15.4 networks are a feasible platform candidate for connecting all health-care-related equipment dispersed across a hospital room to collect critical time-sensitive data about patient health state, such as the heart rate and blood pressure. To meet the quality of service requirements of health-care systems, this paper proposes a multi-priority queue system that differentiates between various types of frames. The effect of the proposed system on the average delay and throughput is explored herein. By employing different contention window parameters, as in IEEE 802.11e, this multi-queue system prioritizes frames on the basis of priority classes. Performance under both saturated and unsaturated traffic conditions was evaluated using a novel analytical model that comprehensively integrates two legacy models for 802.15.4 and 802.11e. To improve the accuracy, our model also accommodates the transmission retries and deferment algorithms that significantly affect the performance of IEEE 802.15.4. The multi-queue scheme is predicted to separate the average delay and throughput of two different classes by up to 48.4% and 46%, respectively, without wasting bandwidth. These outcomes imply that the multi-queue system should be employed in health-care systems for prompt allocation of synchronous channels and faster delivery of urgent information. The simulation results validate these model's predictions with a maximum deviation of 7.6%.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제33권12호
• /
• pp.999-1007
• /
• 2006

역전파 학습 방법은 속도가 느리고, 지역 최소점이나 고원에 빠져 수렴에 실패하는 경우가 많다고 알려져 있다. 이제까지 알려진 역전파의 대체 방법들은 수렴 속도와 변수에 따른 수렴의 안정성 사이에서 불균형이라는 대가를 치루고 있다. 기존의 전통적인 역전파에서 발생하는 위와 같은 문제점 중, 특히 지역 최소점을 탈피하는 기능을 추가하여 적은 저장 공간으로 안정성이 보장되면서도 빠른 수렴속도를 유지하는 알고리즘을 제안한다. 이 방법은 전체 신경망을 은닉층-출력층(hidden to output)을 의미하는 상위 연결(upper connections)과 입력층-은닉층(input to hidden)을 의미하는 하위 연결(lower connections) 2개로 분리하여 번갈아 훈련을 시키는 분리 학습방법을 적용한다. 본 논문에서 제안하는 알고리즘은 다양한 classification 문제에 적용한 실험 결과에서 보듯이 전통적인 역전파 및 기타 개선된 알고리즘에 비해 계산량이 적고, 성능이 매우 좋으며 높은 신뢰성을 보장한다.

• 한국콘텐츠학회논문지
• /
• 제12권4호
• /
• pp.476-484
• /
• 2012

조선 역대 왕의 가계에 등장하는 인물로 구성된 네트워크를 구축하고 분석한 결과, 일반적인 사회 네트워크와 같은 척도 없는 네트워크를 보여주고 있다. 조선왕조 가계 인물 네트워크가 비록 척도 없는 네트워크이지만 네트워크의 지름이 다른 사회 네트워크에 비해 비교적 큰데, 왕조 가계 인물 네트워크는 한 왕에서 다음 왕으로 이어지는 연속적인 특성이 반영된 것이다. K-코어 알고리즘을 도입하여 복잡한 네트워크를 단순화시킬 경우, 복잡한 네트워크에서는 발견하지 못하는 숨겨진 정보를 얻을 수 있는데, 왕조 가계 네트워크에서는 특별한 정보를 얻지 못하였다. 비교적 네트워크의 지름이 크고 길게 이어지는 네트워크에는 k-코어 알고리즘이 적합하지 못함을 의미한다. 단순한 네트워크 구축을 위해 가계 인물 네트워크를 구성하고 있는 소단위 네트워크 즉, 황후, 후궁, 공주나 옹주, 대군이나 군 중심의 네트워크를 구축하여 단순화시키고 그로부터 유용한 정보를 얻고자 하였다. 본 연구에서 복잡한 네트워크의 경우, 데이터베이스에서 분류 가능한 소단위 네트워크를 구축하여 유용한 정보를 도출하는 것도 복잡한 네트워크를 단순화하여 유용한 정보를 도출하는 방법이 될 수 있음을 제시한다. 동시에 역사적인 사실의 정보를 네트워크 관점에서 얻을 수 있음을 본 연구는 제시하고 있다.