• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2005.11a
• /
• pp.191-194
• /
• 2005

본 논문에서는 방송 영상에서 조명효과와 크기변화 등에 강인한 얼굴패턴 검출기법을 제시한다. 제안된 얼굴검출 모델은 영상 전처리 과정과 얼굴패턴 검출 과정으로 이루어진다. 전처리 과정은 조명변화에 대한 보정기능과 다중필터에 의한 후보영역 선별기능으로 구분된다. 얼굴패턴 검출과정은 다단계의 특징지도 생성과정과 패턴분류 과정으로 이루어진다. 특징지도를 생성하기 위하여 가보(Gabor) 필터계층을 포함하는 CNN(Convolutional Neural Networks)모델을 도입하였다. 다양한 배경을 고려한 효과적인 학습을 위하여 본 논문에서는 억제성의 뉴런(Inhibitory neuron)을 포함하는 구조의 CNN모델을 적용한다. CNN으로부터 추출되는 특징집합은 최종 단계에서 WFMM(Weighted Fuzzy Min Max) 모델을 사용하여 분류된다. 이때 사용되는 특징집합의 크기는 분류기의 규모 및 계산량의 결정적인 역할을 준다. 이에 본 연구에서는 최종 분류 과정에 사용되는 특징의 수를 효과적으로 줄이기 위해 FMM모델을 사용하는 적응적인 특징 선별 기법을 제안한다. 또한 실제 영상을 통한 실험결과로부터 제안된 이론의 타당성을 고찰한다.

• The Journal of Korean Medicine
• /
• v.21 no.2
• /
• pp.25-36
• /
• 2000

Objectives : This study demonstrates the effects of Daejo-hwan on learning and memory impairment induced by L-NAME (75 mg/kg) treatment and on cerebral ischemic damage induced by middle cerebral artery (MCA) occlusion in rats. Methods : Daejo-hwan emulsion (73.3 mg/100 g/l ml) was administered to rats along a timed study schedule. The Moms water maze was used for learning and memory test of the rats. The MCA was occluded by using the intraluminal thread method. The brain slices were stained by 2 % triphenyl tetrazolium chloride (TTC) and 1 % cresyl violet solution. Infarct size, neuron cell number and size in penumbra was measured by using computer image analysis system. Results : 1. The escape latency of the Daejo-hwan treated group decreased significantly with respect to the control group. 2.The memory score of the Daejo-hwan treated group showed increase tendency, And the swimming distance was not different between the normal, the control, and the Daejo-hwan treated group. 3. The infarct size of the Daejo-hwan treated group decreased significantly with respect to the control group. 4. The total infarct volume of the Daejo-hwan treated group showed decrease tendency. And the brain edema index of the Daejo-hwan treated group decreased significantly with respect to the control group. 5. The neuron cell number and cell size in penumbra of the Daejo-hwan treated group increased significantly with respect to the control group. Conclusions : According to the above results, it is supposed that Daejo-hwan is clinically applicable to the vascular dementia.

• Journal of KIISE:Software and Applications
• /
• v.29 no.6
• /
• pp.386-395
• /
• 2002

In this paper, we suggest dynamic genetic tree-maps(DGTM) using optimal features on recognizing data. The DGTM uses the genetic algorithm about the importance of features rarely considerable on conventional neural networks and introduces GTM(genetic tree-maps) using tree structure according of the priority of features. Hence, we propose the extended formula, DGTM(dynamic GTM) has dynamic functions to separate and merge the neuron of neural network along the similarity of features.

• 제어로봇시스템학회:학술대회논문집
• /
• 1997.10a
• /
• pp.1601-1604
• /
• 1997

The most important topic in research of mobile robot is path planning in order to avoid with obstacle. In this study the path planning algorithm using a neural network model is proposed. The inputs of neural network are range data which are acquired form laser range finderm and weights are based on difference with goal direction. The thresholds are made by consdiering the marginal distance between mobile robot and obstacle. Consequently the outputs are obtained by multiplying input and weight. The obtained heading directiion enables the mobile robot to approach the goal, without any collision with obstacles around. The effectiveness of the this method of real-time navigation of a mobile robot is estimated by computer simulation in complex environment.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.1E
• /
• pp.3-10
• /
• 2003

This paper describes two kinds of neural networks for stereoscopic vision, which have been applied to an identification of human speech. In speech recognition based on the stereoscopic vision neural networks (SVNN), the similarities are first obtained by comparing input vocal signals with standard models. They are then given to a dynamic process in which both competitive and cooperative processes are conducted among neighboring similarities. Through the dynamic processes, only one winner neuron is finally detected. In a comparative study, with, the average phoneme recognition accuracy on the two-layered SVNN was 7.7% higher than the Hidden Markov Model (HMM) recognizer with the structure of a single mixture and three states, and the three-layered was 6.6% higher. Therefore, it was noticed that SVNN outperformed the existing HMM recognizer in phoneme recognition.

• Proceedings of the KIEE Conference
• /
• 1990.11a
• /
• pp.432-435
• /
• 1990

Neural networks are a class of systems that have many simple processors (neurons) which are highly interconnected. The function of each neuron is simple, and the behavior is determined predominately by the set of interconnections. Thus, a neural network is a special form of parallel computer. Although major impetus for using neural networks is that they may be able to "learn" the solution to the problem that they are to solve, we argue that another, perhaps even stronger, impetus is that they provide a framework for designing massively parallel machines. The highly interconnected architecture of switching networks suggests similarities to neural networks. Here, we present switching applications in which neural networks can solve the problems efficiently. We also show that a computational advantage can be gained by using nonuniform time delays in the network.

• Journal of Advanced Marine Engineering and Technology
• /
• v.21 no.1
• /
• pp.66-70
• /
• 1997

In this paper a new learning algorithm for curvature smoothing and improved generalization for multi-layer neural networks is proposed. To enhance the generalization ability a constraint term of hidden neuron activations is added to the conventional output error, which gives the curvature smoothing characteristics to multi-layer neural networks. When the total cost consisted of the output error and hidden error is minimized by gradient-descent methods, the additional descent term gives not only the Hebbian learning but also the synaptic weight decay. Therefore it incorporates error back-propagation, Hebbian, and weight decay, and additional computational requirements to the standard error back-propagation is negligible. From the computer simulation of the time series prediction with Santafe competition data it is shown that the proposed learning algorithm gives much better generalization performance.

• 제어로봇시스템학회:학술대회논문집
• /
• 1993.10a
• /
• pp.853-858
• /
• 1993

The error back-propagation(BP) algorithm is widely used for finding optimum weights of multi-layer neural networks. However, the critical drawback of the BP algorithm is its slow convergence of error. The major reason for this slow convergence is the premature saturation which is a phenomenon that the error of a neural network stays almost constant for some period time during learning. An inappropriate selections of initial weights cause each neuron to be trapped in the premature saturation state, which brings in slow convergence speed of the multi-layer neural network. In this paper, to overcome the above problem, Micro-Genetic algorithms(.mu.-GAs) which can allow to find the near-optimal values, are used to select the proper weights and slopes of activation function of neurons. The effectiveness of the proposed algorithms will be demonstrated by some computer simulations of two d.o.f planar robot manipulator.

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10b
• /
• pp.302-305
• /
• 1996

This paper presents a visual servoing combined by Neural Network with optimal structure and predictive control for robotic manipulators to tracking or grasping of the moving object. Using the four feature image information from CCD camera attached to end-effector of RV-M2 robot manipulator having 5 dof, we want to predict the updated position of the object. The Kalman filter is used to estimate the motion parameters, namely the state vector of the moving object in successive image frames, and using the multi layer feedforward neural network that permits the connection of other layers, evolutionary programming(EP) that search the structure and weight of the neural network, and evolution strategies(ES) which training the weight of neuron, we optimized the net structure of control scheme. The validity and effectiveness of the proposed control scheme and predictive control of moving object will be verified by computer simulation.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.2E
• /
• pp.49-55
• /
• 2003

The modified neural net pattern recognition equations were attempted to apply to speech recognition. The proposed method has a dynamic process of self-organization that has been proved to be successful in recognizing a depth perception in stereoscopic vision. This study has shown that the process has also been useful in recognizing human speech. In the processing, input vocal signals are first compared with standard models to measure similarities that are then given to a process of self-organization in neural net equations. The competitive and cooperative processes are conducted among neighboring input similarities, so that only one winner neuron is finally detected. In a comparative study, it showed that the proposed neural networks outperformed the conventional HMM speech recognizer under the same conditions.