• KIISE Transactions on Computing Practices
• /
• v.24 no.1
• /
• pp.53-57
• /
• 2018

A poisoning accident is often caused by a situation in which people eat poisonous mushrooms because they cannot distinguish between edible mushrooms and poisonous mushrooms. In this paper, we propose an automatic mushroom recognition system by using the convolutional neural network. We collected 1478 mushroom images of 38 species using image crawling, and used the dataset for learning the convolutional neural network. A comparison experiment using AlexNet, VGGNet, and GoogLeNet was performed using the collected datasets, and a comparison experiment using a class number expansion and a fine-tuning technique for transfer learning were performed. As a result of our experiment, we achieve 82.63% top-1 accuracy and 96.84% top-5 accuracy on test set of our dataset.

• Proceedings of the Korean Society for Rock Mechanics Conference
• /
• 2008.10a
• /
• pp.43-51
• /
• 2008

Currently an increasing number of urban tunnels with small overburden are excavated according to the principle of the New Austrian Tunneling Method (NATM). For rational management of tunnels from planning to construction and maintenance stages, prediction, control and monitoring of displacements of and around the tunnel have to be performed with high accuracy. Computational method tools, such as finite element method, have been and are indispensable tool for tunnel engineers for many years. It is, however, a commonly acknowledged fact that determination of input parameters, especially material properties exhibiting nonlinear stress-strain relationship, is not an easy task even for an experienced engineer. Use and application of the acquired tunnel information is important for prediction accuracy and improvement of tunnel behavior on construction. Artificial Neural Network (ANN) model is a form of artificial intelligence that attempts to mimic behavior of human brain and nervous system. The main objective of this paper is to perform the deformation analysis in NATM tunnel by means of numerical simulation and artificial neural network (ANN) with field database. Developed ANN model can achieve a high level of prediction accuracy.

• Proceedings of the KIEE Conference
• /
• 1999.11c
• /
• pp.649-651
• /
• 1999

This paper presents SNMP remote management system for power plants using micro-controller. Control board supporting LAN-based remote monitor and control is produced and MMI program on PC is also constructed. To expand into Internet, SNMP(Simple Network Management Protocol), the standard network management protocol of TCP/IP protocol suite, is ported to this control board consisted of micro-controller, 80c196KC, and LAN controller, Am7990DC. To overcome the constraints of CPU performance and memory capacity, TCP/P protocol suite is simplified and informations needed to management were implemented in accordance with MIB(Management Information Base) specified in RFC. Also monitoring software is constructed by Visual Basic.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.27 no.3
• /
• pp.137-144
• /
• 1990

We describe how single layer perceptrons and new nonsymmetry feedback type neural networks can be implemented by VLSI CMOS technology. The network described provides a flexible tool for evaluation of boolean expressions and arithmetic equations. About 50 CMOS VLSI chips with an architecture based on two neural networks have been designed and me being fabricated by 2-micrometer double metal design rules. These chips have been developed to study the potential of neural network models for the use in character recognition and for a neural compute.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.10c
• /
• pp.793-795
• /
• 2001

Ad-hoc 망은 인터넷과는 다른 독립적인 망을 구성하는 경우 사용되게 된다. 이러한 Ad-hoc의 특정 호스트가 인터넷 망과의 연결을 시도하는 경우, 이 호스트로의 패킷의 라우팅을 위해서는 모든 경로상의 라우터에서 이 호스트에 대한 정보의 엔트리가 추가되어야만 한다. 이는 망 구현 과정에서 확장성 (Scalability) 문제를 야기하게 된다. 즉, 라우팅 테이블 엔트리의 증가로 차지하는 라우터의 메모리 문제나, 해당 호스트의 정보를 찾고자 할 때의 소요시간으로 적용하기 어렵다. 이러한 확장성 문제를 제거하는 방법으로 자동으로 주소를 설정 하는 방식인 DHCP(Dynamic Host Configuration Protocol)에 대해서 알아 보았다. 그리고. DHCP의 문제점인 기존의 라우팅 정보를 손실되는 문제가 있어 적용하기 어렵다. 이에 본 논문은 확장성 문제를 제거하는 동시에 엑세스 포인트(access point)에서 NAT(Network Address Translation) 기술을 적용하여 AODV(Ad-Hoc On-Demand Distance Vector) 라우팅 프로토롤과 인터넷의 연동을 위한 메커니즘을 제안한다.

• Proceedings of the KIEE Conference
• /
• 2002.07d
• /
• pp.2158-2160
• /
• 2002

In this paper, we designed the adaptive fuzzy controller(AFLC) using neural network and tabu search. We tuned the weights of neural network changing adaptively input/output gain of fuzzy logic controller and the gain of fuzzy logic controller using tabu search. To evaluate the proposed method's effectiveness, we apply the proposed AFLC to the speed control of an actual AC servomotor system. The experimental results show that AFLC has the better control performance than PI controller in terms of settling time, rising time and overshoot.

• Proceedings of the KIEE Conference
• /
• 1998.07c
• /
• pp.1046-1049
• /
• 1998

This paper deals with the new fault detection technique for transmission line using Neuro-fuzzy Scheme. Neuro-fuzzy Scheme is ANFIS(Adaptive-network Fuzzy Inference System) based on fusion of fuzzy logic and neural networks. The proposed scheme has five layers. Each layer is the component of fuzzy Inference system and performs different action. Using learning method of neural network, fuzzy premise and consequent parameters is tuned properly.

• Journal of KIISE:Information Networking
• /
• v.35 no.4
• /
• pp.268-281
• /
• 2008

It is difficult to construct streaming services based on the overlay networks without any loss of scalability. DONet is one of the most representative streaming overlay network protocols without managing any specific structure. Since DONet does not support the nodes on private networks, it can be considered that the performance of the overlay is not the best. Hole Punching is one of the famous techniques participating the nodes on private networks to streaming overlay networks by using a rendezvous server. However, using only a single rendezvous server cannot be suggested in P2P environment, because it can cause problems in terms of scalability and so on. In this paper, we propose DONet-p, an extension of DONet with Distributed Hole Punching techniques. It supports the nodes on private networks without toss of scalability. The experimental results show the better performance and scalability than DONet with a minimum overhead for additional control messages.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.45 no.8
• /
• pp.114-120
• /
• 2008

In this paper, we extract the parasitic elements of the metal-insulate-metal(MIM) capacitor using short-open calibration (SOC). The scattering matrixes of short, open, and MIM structures in strip lines are measured by full electro-magnetic (EM) simulator and vector network analyser. The full EM simulations are performed by finite element method (FEM) that was fitted three dimensional structure analysis. The electro-magnetic effects of MIM capacitor laminated in the multi-layered structures are proposed the II equivalent circuit with lumped elements, and the relations between the measured scattering parameters of the MIM structures and lumped elements in the circuits are shown by performing 2 port network analysis. The extracted lumped elements using the proposed SOC method are independent to frequencies.

• Smart Media Journal
• /
• v.4 no.3
• /
• pp.38-43
• /
• 2015

The increased number of smart devices leads users to share multimedia contents with others. Smart device users transmit multimedia files using apps. For multimedia file transfer under NAT (Network Address Translator), a lot of apps exploit a server with public IP or short range communication technologies such as WiFi Direct or Bluetooth because smart devices with private IP address cannot communicate each other directly. We first proposed the push service based file transmission scheme without an additional server and distance limitation. Second, for the push service based transmission, we proposed congestion control scheme to adapt transmission rate to the receiver's link bandwidth. Third, we considered bandwidth aware transmission for senders with different bandwidth. We implemented apps to evaluate the proposed scheme, and conducted experiments with smart devices.