• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.18 no.10
• /
• pp.1422-1432
• /
• 1993

Nonlinear mapping function of the HCNN( Hidden Control Neural Network ) can change over time to model the temporal variability of a speech signal by combining the nonlinear prediction of conventional neural networks with the segmentation capability of HMM. We have two things in this paper. first, we showed that the performance of the HCNN is better than that of HMM. Second, the HCNN with its prediction error measure given by weighted distance is proposed to use suitable distance measure for the HCNN, and then we showed that the superiority of the proposed system for speaker-independent speech recognition tasks. Weighted distance considers the differences between the variances of each component of the feature vector extraced from the speech data. Speaker-independent Korean digit recognition experiment showed that the recognition rate of 95%was obtained for the HCNN with Euclidean distance. This result is 1.28% higher than HMM, and shows that the HCNN which models the dynamical system is superior to HMM which is based on the statistical restrictions. And we obtained 97.35% for the HCNN with weighted distance, which is 2.35% better than the HCNN with Euclidean distance. The reason why the HCNN with weighted distance shows better performance is as follows : it reduces the variations of the recognition error rate over different speakers by increasing the recognition rate for the speakers who have many misclassified utterances. So we can conclude that the HCNN with weighted distance is more suit-able for speaker-independent speech recognition tasks.

• Spatial Information Research
• /
• v.20 no.5
• /
• pp.77-89
• /
• 2012

Walking is the most basic personal mobility and its importance and concern is ever increasing with the highlighting of a new paradigm, such as transit oriented development, sustainable development and revitalization of green transport. The existing analytical research on pedestrian network is using a pedestrian's moving distance to a destination and integration in space syntax theory as its representative accessibility factors. However, the uniplanar network moving distance fails to reflect topographic characteristics, so the moving distance could show a similar result value in case of the regions for analysis that have a similar network structure to each other. Accordingly, the aim of this study is to suggest a new analytical methodology on pedestrian network accessibility in consideration of the grade in pedestrian sections and a pedestrian's size. this study, in its analysis of a uniplanar pedestrian network moving distance, analyzed the pedestrian network moving distance in consideration of the grade in pedestrian sections, and even the pedestrian network moving distance in consideration of a pedestrian's size, and suggested the methodology on pedestrian network accessibility analysis in consideration of a more substantive pedestrian's characteristics. It is hoped that the methodology used by this study will be used as the methodology on pedestrian network analysis which can reflect topographic characteristics in the pedestrian network analysis, and take a more substantive pedestrian's movement into account.

• Proceedings of the KAIS Fall Conference
• /
• 2005.05a
• /
• pp.180-183
• /
• 2005

본 논문에서는 인쇄체 아라비아 숫자를 인식하기 위해 가중 템플릿 정합 방법을 제안한다. 가중 템플릿 정합은 패턴의 특징이 나타나는 영역에 해밍거리(Hamming Distance) 의 가중치를 두어 패턴 특징을 강조하여 숫자 패턴의 인식률을 높이는 것이다. 또한 패턴의 표면을 울퉁불퉁한 영상으로 만드는 한 두 픽셀의 랜덤 노이즈를 제거하기 위하여 본 연구에서는 트리밍(trimming) 기법을 적용하였다. 실험에서는 트리밍을 하지 않고 단순 템플릿 정합을 사용했을 때의 혼돈 행렬(confusion matrix)과 트리밍을 한 후 가중 템플릿 정합을 사용했을 때 혼돈 행렬을 서로 비교해 인식률이 크게 향상된 것을 보인다.

• The Journal of the Acoustical Society of Korea
• /
• v.15 no.5
• /
• pp.61-64
• /
• 1996

To improve the performance of a recognition system, namely the recognition rate, we propose a hew implementation of perceptual distance using LPC cepstrum(perceptual cepstrum, PLC). The PLC is caculated by convolution of a usual LPC cepstrum and a perceptual lifter(PL). To caculate PL, we define a new weighting function in the linear frequency domain considering the frequency scale(Bark-scale) characteristics. The PL is the inverse Fourier transform of the exponents of the weighting function. We verified our method through the speech recognition experiments. The performance of PLC was compared with that of the rasied sine liftering method.

• Journal of Korea Water Resources Association
• /
• v.42 no.2
• /
• pp.131-139
• /
• 2009

The weather accidents by global warming effect are increasing rapidly whole world. Flood forcasting system and hydrological database are operated by almost all the countries in the world. An objective of this study is to research revised methods of missing rainfall data and find more effective revised method for this operating system. 194 rainfall data of the Han river basin is used. Arithmetic average method, coefficient of correlation weighting method and inverse distance weighting method are compared to estimate revised methods. The result from the analysis shows that coefficient of correlation weighting method is best quantitatively among the 3 methods.

• The Journal of the Acoustical Society of Korea
• /
• v.15 no.4
• /
• pp.36-42
• /
• 1996

In this paper, we have proposed the robust estimation of HMM parameters which is based on CDDN(codeword dependent distance normalization)and codeword weighting by distance. The proposed method has used a distance normalization based on the characteristics of a codeword dependent distribution and have computed fuzzy contributions of codeword to a input vector with a fuzzy objective function. From experimental results, we have shown the effectiveness of the proposed method in that the correction rate of the proposed method is improved 4.5% over the conventional FVQ based method. Especially, the application of distance weighting to smoothing of output probability is improved the performance of 2.5% compared to distance based codeword weighting.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.32 no.2D
• /
• pp.95-101
• /
• 2012

Layout Criteria are developed for an access mode's on and off facility at multiple transfer centers in this paper. Layout Criteria are location issues between a main mode and access modes in terms of the on and off facility one another. The total distance between them has been suggested to be minimized. In the distance calculation stairs are considered to be more difficult and than open space. On the other hand an escalator and an elevator are treated as easier than open space. Considering the number of people between on and off facility, the weighted average distance is suggested to be a MOE(Measure of effectiveness) for layout criteria at multiple transfer centers. Finally, the layout criteria are applied to the existing Kimpo airport terminal and some improvement ideas are suggested.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1998.06c
• /
• pp.51-54
• /
• 1998

본 논문에서는 자동차 잡음 환경에서 녹음된 데이터 베이스를 이용하여 인식 시스템의 성능을 향상시키기 위한 효율적인 잡음 제거 방법을 연구하였다. 먼저, 잡음 및 주변 환경 변화에 강인한 것으로 알려져 있는 특징 벡터들의 인식 성능을 비교하교, 가중 켑스트랄 거리 측정 방법을 이용한 인식 실험을 통하여 시스템의 성능 향상을 확인하였다. 실험 결과, 본 논문에서 기준 시스템으로 사용한 LPC 켑스트럼의 경우에 비하여 MFCC나 root-cepstrum을 사용한 경우 인식률이 향상되었다. 켑스트럼간의 거리 측정에 있어서는 RPS와 BPL과 같은 가중 켑스트랄 거리 측정 함수들이 인식 성능 향상에 도움을 주었다. 또한 켑스트럼 평균 차감법이라는 간단한 잡음 제거기술을 적용하여 자동차 잡음 환경에서 인식 성능 향상을 보였다. 마지막으로, 차량 항법용 음성 인식 시스템의 실시간구현을 위하여 여러 경우의 인식 성능을 비교하고, 메모리 량과 실행 시간 등을 고려하여 최적 시스템을 제시하였다.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.2 no.4
• /
• pp.175-182
• /
• 2000

Spatial interpolation of daily temperature forecasts and observations issued by public weather services is frequently required to make them applicable to agricultural activities and modeling tasks. In contrast to the long term averages like monthly normals, terrain effects are not considered in most spatial interpolations for short term temperatures. This may cause erroneous results in mountainous regions where the observation network hardly covers full features of the complicated terrain. We developed a spatial interpolation model for daily minimum temperature which combines inverse distance squared weighting and elevation difference correction. This model uses a time dependent function for 'mountain slope lapse rate', which can be derived from regression analyses of the station observations with respect to the geographical and topographical features of the surroundings including the station elevation. We applied this model to interpolation of daily minimum temperature over the mountainous Korean Peninsula using 63 standard weather station data. For the first step, a primitive temperature surface was interpolated by inverse distance squared weighting of the 63 point data. Next, a virtual elevation surface was reconstructed by spatially interpolating the 63 station elevation data and subtracted from the elevation surface of a digital elevation model with 1 km grid spacing to obtain the elevation difference at each grid cell. Final estimates of daily minimum temperature at all the grid cells were obtained by applying the calculated daily lapse rate to the elevation difference and adjusting the inverse distance weighted estimates. Independent, measured data sets from 267 automated weather station locations were used to calculate the estimation errors on 12 dates, randomly selected one for each month in 1999. Analysis of 3 terms of estimation errors (mean error, mean absolute error, and root mean squared error) indicates a substantial improvement over the inverse distance squared weighting.

• Journal of Digital Convergence
• /
• v.12 no.8
• /
• pp.321-327
• /
• 2014

Feature selection is the technique to improve the classification performance by using a minimal set by removing features that are not related with each other and characterized by redundancy. This study proposed new feature selection using the distance between the center of gravity of the bounded sum of weighted fuzzy membership functions (BSWFMs) provided by the neural network with weighted fuzzy membership functions (NEWFM) in order to improve the classification performance. The distance-based feature selection selects the minimum features by removing the worst features with the shortest distance between the center of gravity of BSWFMs from the 24 initial features one by one, and then 22 minimum features are selected with the highest performance result. The proposed methodology shows that sensitivity, specificity, and accuracy are 97.7%, 99.7%, and 98.7% with 22 minimum features, respectively.