• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.31B no.10
• /
• pp.123-129
• /
• 1994

The segmentation of range image is essential to recognize the three dimensional object. Generally, surface curvature is well-known feature for segmentation and classification of the fange image, but it is sensitive to noies. In this paper, we propose the structure of hierarchical neural network using surface curvature for segmentation of range images. The hierarchical structure of neural networks is robust to noise and the result of segmentaion is better than conventional optimization method of single level.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.48 no.1
• /
• pp.18-23
• /
• 2011

We have investigated the Intelligent Shift-map Module(ISM) to improve some problems in the conventional Automatic Transmission(AT) for automobiles. The typical AT lacks flexibility regarding the shift point because it does not consider the driver's habits and inclinations. Also it often is occurred phenomenon like kick-down. Therefore, we designed a decision module which considers the driving style of the individual driver. The driving style was determined by the inclination of the driver and the driving technique using actual automobile data. The Hierarchical Neural Network(HNN) was applied in generating an intelligent shift map with Multilayer Neural Network(MNN). It was found that the proposed ISM provided a suitable shift point and time because the necessary toque and velocity of the automobile was considered along with the driving style of each driver when designing the ISM.

• Annual Conference on Human and Language Technology
• /
• 1991.10a
• /
• pp.287-299
• /
• 1991

본 논문은 하드웨어 구현이 가능한 신경 회로망을 구성하여 한글 문자 인식을 수행하였다. 먼저 입력 장치로부터 받아들인 문자 영상은 인식 속도를 높히기 위하여 특별한 전처리 과정 없이 직접 피쳐를 추출하였으며 추출한 피쳐로는 하드웨어 구현이 용이한 교차 피쳐와 투영 피쳐를 이진화로 코딩하였다. 신경 회로망의 하드웨어 구현을 가능하게 하기위해서 정수형 연결 강도와 비선형 Hard-limit 함수를 가지고 학습을 하는 Rounding 학습 방법을 도입하여 학습시켰으며 한글의 구조적 특성을 이용하여 한글을 유형별로 Module화 및 Submodule화 작업을 수행한 다음 인식하는 계층적인 문자 인식 시스템을 구성하였다. 그리고 이러한 방법을 이용하여 한글 문자 인식용 CMOS 신경회로망 Chip을 설계하였다.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 1995.10a
• /
• pp.280-283
• /
• 1995

계층형 신경회로망은 학습능력이나 비선형사상능력을 가지고 있고, 그 특징을 이용하여 패턴인식이나 동정 및 제어 등에의 적용이 시도되어 성과를 올리고 있다. 현재, 그 학습법으로 널리 이용되고 있는 것이 역전파학습법으로 최급 강하법이나 공액경사법 등의 최적화 방법이 적용되고 있지만, 학습에 많은 시간이 걸리는 점, 국소적 최적해(local minima)에 해의 수렴이 이루어져 오차가 충분히 작게 되지 않는 점 등이 문제점으로 지적되고 있다. 본 논문에서는 Hu에 의해 고안된 random 탐색법과 조합된 random tabu 탐색법으로 최적결합계수를 구하는 학습알고리즘으로, 국소적 최적해에 수렴하는 것을 방지하고, 수렴정도를 개선하는 새로운 방법을 이용하여 회전기계의 이상진동진단에 적용가능성을 검토하고 오차역전파법에 의한 진단결과와 비교검토한다.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.36C no.11
• /
• pp.84-91
• /
• 1999

In this paper, we could analyze the comparison with applied WFM to the quantization triangularity fuzzy function and triangularity Fuzzy function. In order to improve on a fault which not remove completely noise included image according to a peculiarity of noise, we got to realize FDNN of the high speed weight eliminating noise included image, minimizing the lost of information, obtaining information of suitability owing to applied Fuzzy Algorithm to DBNN of a hierarchical structure. We could analyze the comparison with a power of WFM and FDNN using simulation We could find to superiority the proposed FDNN )n a result which was the comparison of MSE for the boats image.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.5
• /
• pp.1130-1137
• /
• 1996

Every neural network has some kind of feedback. For the sake of analyzing fundamental aspects of information processing in neural nets, a net without feedbacks is an important theoretical model. But here we focus on a recurrent neural net which delay synapses as a realistic dynamical model of nervous systems. Synaptic connections are determined by a version of the Hebb rule (correlation type rule). We use a statistical neurodynamic method to explain the retrieval dynamics of the network. The result of the analysis for the sequential associativ e memory with delay synapses is compared with computer simulation. We have succeeded in explaining the dynamics of this network by theoretical analyses.

• Proceedings of the KIEE Conference
• /
• 2001.11c
• /
• pp.123-126
• /
• 2001

종합적인 공정 제어 자동화 시스템 구현을 위해 계층적이고 개방형 방식에 의한 시스템 구축이 이루어지고 있다. 상위 계층 시스템은 하위 계층 제어 시스템의 제어기 설정치를 결정하는 방법으로 다양한 의사결정(Decision Making)방법을 도입하여 하위 계층 시스템과 연계하여 계층적인 종합 공정 제어 자동화 시스템 구축을 시도하고 있다. 본 연구에서는 상위 계층 시스템 구현을 위해 신경회로망 방식을 채택한 상위 지능형 제어 시스템을 제안하여 연속형 프로세스의 최적 의사 결정을 효과적으로 할 수 있도록 하였고 이를 실현화 하는데 있어 UML방식의 객체지향 설계방식을 도입함으로써 시스템의 재 사용성 및 확장성을 가지는 개방형 상위 의사 결정 시스템을 개발하였다. 개발된 시스템을 수처리 연속 공정인 약품주입 공정에 적용하였다.

• Proceedings of the Korea Association of Information Systems Conference
• /
• 2000.11a
• /
• pp.1.2-24
• /
• 2000

The field of neural networks has been a recent surge in activity as a result of progress in developments of efficient training algorithms. For this reason, and coupled with the widespread availability of powerful personal computer hardware for running simulations of networks, there is increasing focus on the potential benefits this field can offer. The neural network may be viewed as an advanced pattern recognition technique and can be applied in many areas such as financial time series forecasting, medical diagnostic expert system and etc.. The intention of this study is to build and implement one simple artificial neural networks hereinafter called ANN. For this purpose, some literature survey was undertaken to understand the structures and algorithms of ANN theoretically. Based on the review of theories about ANN, the system adopted 3-layer back propagation algorithms as its learning algorithm to simulate one case of medical diagnostic model. The adopted ANN algorithm was performed in PC by using turbo PASCAL and many input parameters such as the numbers of layers, the numbers of nodes, the number of cycles for learning, learning rate and momentum term. The system output more or less successful results which nearly agree with goals we assumed. However, the system has some limitations such as the simplicity of the programming structure and the range of parameters it can dealing with. But, this study is useful for understanding general algorithms and applications of ANN system and can be expanded for further refinement for more complex ANN algorithms.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.13 no.1
• /
• pp.69-76
• /
• 1999

This paper proposed a design of intelligent supervisory control systems for maintenance of highway traffic information equiprrent and processing algorithm of equiprrent fault data. The fault data of highway traffic equipment are transmitted from rerrnte supervisory controller to central supervisory system by real time, the transmitted fault data are anaIyzed the characteristic using evaluation algorithm of fault data in central supervisory system. The evaluation algorithm includes a neural network and fault knowlOOge-base for processing the multi-generated fault data. For validating the evaluation algorithm of intelligent supervisory control systems, the rrethod of analysis used to the five pattern of binary signal by transmitted real time and the opTclting user-interface constructed in central supervisory system.

• Journal of Institute of Control, Robotics and Systems
• /
• v.4 no.4
• /
• pp.499-505
• /
• 1998

In developing an automated surface inspect algorithm, we have designed a hierarchical classifier using neural network. The defects which exist on the surface of cold mill strip have a scattering or singular distribution. We have considered three major problems, that is preprocessing, feature extraction and defect classification. In preprocessing, Top-hit transform, adaptive thresholding, thinning and noise rejection are used Especially, Top-hit transform using local minimax operation diminishes the effect of bad lighting. In feature extraction, geometric, moment, co-occurrence matrix, and histogram ratio features are calculated. The histogram ratio feature is taken from the gray-level image. For defect classification, we suggest a hierarchical structure of which nodes are multilayer neural network classifiers. The proposed algorithm reduced error rate by comparing to one-stage structure.