• Proceedings of the Korean Information Science Society Conference
• /
• 2003.04d
• /
• pp.187-189
• /
• 2003

웨이블릿 변환에 기반한 동영상 코덱에서의 움직임 예측 기법은 OCT 기반 코덱과 유사하게 이전 프레임과의 움직임 예측을 통하여 수행된다. 그러나, 현재 프레임이 이전 프레임을 참조하므로 네트워크상의 전송시 이전 프레임에 발생한 오류가 전달되는 오류 전파의 문제도 발생하게 된다. 본 논문에서는 웨이블릿 변환된 프레임의 특성을 이용하여 최상위 레벨의 LL 부대역만 이전 프레임과의 움직임 예측을 수행하고, 나머지 부대역에 대하여 프레임 내의 상위레벨의 부대역이 하위 부대역을 창조하여 예측 및 보상을 수행하여 오류전파의 가능성을 최소화하는 Inter-Intra ME 동영상 코덱을 제안한다 제안된 움직임 예측을 사용하여 MAD(Mean-Absolute Differences)를 측정한 결과, 프레임간 변화가 심한 경우에는 제안된 기법과 이전 프레임의 부대역을 참조한 기법 사이의 압축율은 유사하게 나타났으며, 변화가 적은 경우에는 이전 프레임을 참조하는 것의 압축율이 높게 나타났다. 그러나, 네트워크 전송시 발생하는 오류전파에는 제안된 기법의 성능이 우수한 것으로 나타났다.

• Proceedings of the Korea Society for Energy Engineering kosee Conference
• /
• 1994.11a
• /
• pp.17-22
• /
• 1994

오류역전파 신경망을 인산형 연료전지의 조업변수인 산소 및 수소 유량, 작동온도에 대하여 학습시켜 연료전지 모델을 구성하였다. 또한 구성된 모델을 이용하여 다양한 조업조건에서의 단위전지 성능을 예측하여 이를 실험결과와 비교하였으며, 학습된 신경망을 ASPEN PLUS의 단위공정으로 도입하여 50kW 출력의 연료전지 공정을 구성한 후 조업변수에 대한 영향을 살펴보았다. 3개의 층으로 구성된 오류역전파 신경망은 학습단계상수와 모멘텀이 각각 0.7 및 0.9인 경우 단위전지 성능곡선을 가장 정확히 학습하였으며, 이에 의하여 구성된 신경망 모델은 수소 및 산소의 유량, 온도의 변화에 따른 단위전지 성능곡선의 변화를 정확히 예측하였다. 연료전지 전체공정의 모사에서는 개질기의 경우 $600^{\circ}C$의 상압에서 수증기/탄화수소 비율이 2.6일 때, 연료전지의 경우 작동온도가 190~20$0^{\circ}C$일 때 연료전지의 출력이 최대값을 나타내었으며, 단위전지의 전기화학적 효율은 약 45%, 수소이용률은 약 61%, 발전시스템 전체의 효율은 18%이었다.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.10 no.6
• /
• pp.213-223
• /
• 2006

In this study, artificial neural network is applied to the evaluation of bearing capacity of the PHC auger-drilled piles at sites of domestic decomposed granite soils. For the verification of applicability of error back propagation neural network, a total of 168 data of in-situ test results for PHC auger-drilled plies are used. The results show that the estimation of error back propagation neural network provide a good matching with pile test results by training and these results show the confidence of utilizing the neural networks for evaluation of the bearing capacity of piles.

• Journal of KIISE:Information Networking
• /
• v.30 no.6
• /
• pp.695-704
• /
• 2003

A wavelet-based video stream is more susceptible to the network transmission errors than DCT-based video. This is because bit-errors in a subband of a video frame affect not only the other subbands within the current frame but also the subsequent frames. In this paper, we propose a video source coding scheme called IPC(Intra Prediction Coding) scheme in order to reduce the error propagation to the subsequent frames. In the proposed scheme, a subband except LL subband in the current frame refers to the lower-level subband within the same frame. This reduces the error propagation to subsequent frames. We evaluated the performance of our proposed scheme in the simulated wireless network environment. As a result of tests, it was shown that the proposed algorithm shows better performance than MRME in a heavy motion image sequence while IPC outperforms MRME at a high bit-rate in small motion image sequence.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2004.10a
• /
• pp.153-157
• /
• 2004

본 논문은 패턴인식을 위해 사용할 수 있는 감독학습을 이용한 supervised IAFC neural network 1과 supervised IAFC neural network 2를 제안하였다 Supervised IAFC neural network 1과 supervised IAFC neural network 2는 LVQ(Learning Vector Quantization)를 퍼지화한 새로운 퍼지 학습법칙을 사용하고 있다. 이 새로운 퍼지 학습 법칙은 기존의 학습률 대신에 퍼지화된 학습률을 사용하고 있는데, 이 퍼지화된 학습률은 조건 확률을 퍼지화 한 것에 근간을 두고 있다. Supervised IAFC neural network 1과 supervised IAFC neural network 2의 성능과 오류역전파 신경회로망의 성능을 비교하기 위하여 iris 데이터를 사용하였는데, 실험결과 supervised IAFC neural network 2 의 성능이 오류역전파 신경회로망의 성능보다 우수함이 입증되었다.

• Journal of KIISE:Software and Applications
• /
• v.28 no.9
• /
• pp.603-611
• /
• 2001

This paper proposes a novel fast layer-by-layer algorithm that has better generalization capability. In the proposed algorithm, the weights of the hidden layer are updated by the target vector of the hidden layer obtained by least squares method. The proposed algorithm improves the learning speed that can occur due to the small magnitude of the gradient vector in the hidden layer. This algorithm was tested in a handwritten digits recognition problem. The learning speed of the proposed algorithm was faster than those of error back propagation algorithm and modified error function algorithm, and similar to those of Ooyen's method and layer-by-layer algorithm. Moreover, the simulation results showed that the proposed algorithm had the best generalization capability among them regardless of the number of hidden nodes. The proposed algorithm has the advantages of the learning speed of layer-by-layer algorithm and the generalization capability of error back propagation algorithm and modified error function algorithm.

• Journal of Bio-Environment Control
• /
• v.20 no.4
• /
• pp.277-282
• /
• 2011

We carried out shape recognition. We found out cucumber's feature shape by means of neural network and back propagation algorithm. We developed an algorithm which finds object position and shape in real image and we gained following conclusion as a result. It was processed for feature shape extraction of cucumber to detect automatic. The output pattern rates of the miss-detected objects was 0.1~4.2% in the output pattern which was recognized as cucumber. We were gained output pattern according to image resolution $445{\times}363$, $501{\times}391$, $450{\times}271$, $297{\times}421$. It was appeared that no change was detected. When learning pattern was increased to 25, miss-detection ratio was 16.02%, and when learning pattern had 2 pattern, it didn't detect 8 cucumber in 40 images.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.39 no.1
• /
• pp.23-34
• /
• 2002

This paper proposes a new neuro-fuzzy system that is fast trained by evolution strategy and describes application results of the proposed system to nonlinear system identification to show its usefulness. As training methods of neuro-fuzzy systems, modified error back-propagation algorithms and genetic algorithms have been used so far. However, the former has some drawbacks such as long training time, falling to local optimum, and experimental selecting of learning rates and the latter has difficulty in precise searching solutions because genetic algorithms represents solutions as genotype individuals. The evolution strategy we used can do precise search because its individuals are represented as phenotype real values, it seldom falls into a local optimum, and its training speed is faster than error back-propagation algorithms. We apply our neuro-fuzzy systems to nonlinear system identification. It was found from experiments that training speed is fast and the training results were considerably good.

• The Journal of the Acoustical Society of Korea
• /
• v.21 no.5
• /
• pp.430-437
• /
• 2002

Having a variety of good characteristics against other pattern recognition techniques, the multilayer perceptron (MLP) has been widely used in speech recognition and speaker recognition. But, it is known that the error backpropagation (EBP) algorithm that MLP uses in learning has the defect that requires restricts long learning time, and it restricts severely the applications like speaker recognition and speaker adaptation requiring real time processing. Because the learning data for pattern recognition contain high redundancy, in order to increase the learning speed it is very effective to use the online-based learning methods, which update the weight vector of the MLP by the pattern. A typical online EBP algorithm applies the fixed learning rate for each update of the weight vector. Though a large amount of speedup with the online EBP can be obtained by choosing the appropriate fixed rate, firing the rate leads to the problem that the algorithm cannot respond effectively to different learning phases as the phases change and the number of patterns contributing to learning decreases. To solve this problem, this paper proposes a Changing rate and Omitting patterns in Instant Learning (COIL) method to apply the variable rate and the only patterns necessary to the learning phase when the phases come to change. In this paper, experimentations are conducted for speaker verification and speech recognition, and results are presented to verify the performance of the COIL.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2003.11b
• /
• pp.1173-1176
• /
• 2003

JPEG2000[1][2]은 DWT(Discrete Wavelet Transform)[3] 변환을 기반으로 하는 이미지의 새로운 표준으로 오류가 빈번히 발생하는 네트워크상의 전송을 위한 오류 제어 기법들을 제공한다. JPEG2000에서 제공하는 오류 제어 기법은 스트림 내에서 오류를 발견하고 오류 전파를 제어하지만, 손실된 웨이블릿 계수를 복원하지는 못한다. 본 논문에서는 DWT 변환 후 각 부대역 및 이웃의 상호 연관성을 이용하여 오류로 인해 손실된 웨이블릿 계수를 효과적으로 예측하여 이미지의 화질을 향상시키는 오류 은닉 기법을 제안한다. 제안된 기법의 성능을 무선네트워크 환경에서 모의 실험한 결과 기존의 오류 제어 기법보다 성능이 뛰어났고 특히 이미지 내의 변화가 적은 경우 부대역 및 이웃 웨이블릿 계수의 상호 연관성이 크므로 더욱 효과적인 성능을 보였다.