• The Journal of the Acoustical Society of Korea
• /
• v.29 no.1
• /
• pp.56-61
• /
• 2010

This paper describes optimization methods of acoustic models in HMM-based continuous speech recognition. Most of the conventional speech recognition systems use the same number of Gaussian mixture components for each HMM state. However, since the number of data samples available for each state is different from each other, it is possible to reduce the overall number of model parameters and the computational cost at the decoding step by optimizing the number of Gaussian mixture components. In this study, we introduced the Gaussian mixture weight term at the merging stage of Gaussian components in the minimum description length (MDL) based acoustic modeling optimization. Experimental results showed that the proposed method can obtain better ASR accuracy than the previous optimization method which does not consider the Gaussian mixture weight term.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2002.10a
• /
• pp.549-554
• /
• 2002

In Powder System, SCC demands high dosage of superplasticizer and a lage amout of powder for suitable fluidity and viscosity. Okamura's method of most representative mixing design method in SCC of Powder-System is unfavorable economically because of using a large amount of powder. In addition, many ready-mixed concrete plants do not use his mix design method and procedure due to complexity for practical application. Therefore, Nan Su proposed more simple mix design method than Okamura's. It had an advantage in simplicity in practical application and required a smaller amount of powders compared with Okamura's method. This paper proposed an optimal mixture proportion of SCC with consideration of Nan Su's method. The new and modified mix design method required a smaller amount of powder than that of Nan Su's. To check the properties of SCC, considered with the requirements specified by the Japanese Society of Civil Engineering.(JSCE)

• International Journal of Control, Automation, and Systems
• /
• v.4 no.6
• /
• pp.705-713
• /
• 2006

For the analysis of multiunit extracellular neural signals as multiple spike trains, neural spike sorting is essential. Existing algorithms for the spike sorting have been unsatisfactory when the signal-to-noise ratio(SNR) is low, especially for implementation of fully-automated systems. We present a novel method that shows satisfactory performance even under low SNR, and compare its performance with a recent method based on principal component analysis(PCA) and fuzzy c-means(FCM) clustering algorithm. Our system consists of a spike detector that shows high performance under low SNR, a feature extractor that utilizes projection pursuit based on negentropy maximization, and an unsupervised classifier based on Gaussian mixture model. It is shown that the proposed feature extractor gives better performance compared to the PCA, and the proposed combination of spike detector, feature extraction, and unsupervised classification yields much better performance than the PCA-FCM, in that the realization of fully-automated unsupervised spike sorting becomes more feasible.

• Proceedings of the KSR Conference
• /
• 2011.05a
• /
• pp.1161-1166
• /
• 2011

A camera is mounted vertically downwards viewing the people heads from the top. This configuration is successful in people counting technique especially when only a few isolated people pass through a counting region in a non-crowded situation. Thus, this paper describes object counting which detects and count moving people using mixture of gaussian and motion vector. This method is intended to estimates the number of people in outdoor environment. This method use single gaussian background modeling which is more robust an noise and has adaptiveness. The experimental results that is based on mixture of gaussian and motion vector is also helpful to design intelligent surveillance.

• 한국전산유체공학회:학술대회논문집
• /
• 2009.11a
• /
• pp.177-184
• /
• 2009

Cavitating flow simulation is of practical importance for many engineering systems, such as marine propellers, pump impellers, nozzles, injectors, torpedoes, etc. The present work has focused on the simulation of cavitating flow past cylinders with strong side flows. The governing equation is the Navier-Stokes equation based on the homogeneous mixture model. The momentum and energy equation is in the mixture phase while the continuity equation is solved liquid and vapor phase, separately. An implicit dual time and preconditioning method are employed for computational analysis. For the code validation, the results from the present solver have been compared with experiments and other numerical results. A fairly good agreement with the experimental data and other numerical results have been obtained. After the code validation, the strong side flow was applied to include the wake flow effects of the submarine or ocean tide.

• 한국전산유체공학회:학술대회논문집
• /
• 2009.04a
• /
• pp.149-154
• /
• 2009

The cavitating flow simulation is of practical importance for many engineering systems, such as marine propellers, pump impellers, nozzles, injectors, torpedoes, etc. The present work has focused on the simulation of cavitating flow past cylinders with strong side flow. The governing equation is the Navier-Stokes equation based on homogeneous mixture model. The momentum and energy equation is in the mixture phase while the continuity equation is solved liquid and vapor phase, separately. An implicit dual time and preconditioning method are employed for computational analysis. The results from the present solver have been in a fairly good agreement with the experimental data and other numerical results. After the code validation the strong side flow was applied to include the wake flow effect of the submarine.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.12 no.3
• /
• pp.19-26
• /
• 2004

A prior fundamental study was executed using a constant volume chamber to improve the burning properties of lean pre-mixture by the injection of active radicals generated in the sub-chamber. In consequence, RI method shows remarkable progress in the aspects of burning velocity and combustible lean limit compared with SI method. In this study, the necessary additional works have been performed to be based on the former results. We changed parameters as the initial temperature and the initial pressure of mixture. And the effects of residual gas at issue in a real engine were investigated. As a result, the effects of initial temperature were significant, but on the other hand, those of initial pressure were slight. The correlation of passage hole number between overall passage hole area was grasped. And the more detailed analysis is required on residual gas.

• International Journal of Contents
• /
• v.9 no.1
• /
• pp.1-5
• /
• 2013

Standard Gaussian Mixture Model (GMM) is a well-known method for image segmentation. However, one of its problems is that we consider the pixel as independent to each other, which can cause the segmentation results sensitive to noise. It explains why some of existing algorithms still cannot segment texts from the background clearly. Therefore, we present a new method in which we incorporate the spatial relationship between a pixel and its neighbors inside $3{\times}3$ windows to segment the text. Our approach works well with images containing texts, which has different sizes, shapes or colors in case of light changes or complex background. Experimental results demonstrate the robustness, accuracy and effectiveness of the proposed model in image segmentation compared to other methods.

• The Korean Journal of Applied Statistics
• /
• v.13 no.2
• /
• pp.505-514
• /
• 2000

A Markov Chain Monte Carlo method with data augmentation is developed to compute the features of the posterior distribution. For each observed failure epoch, we introduced mixture failure model of Rayleigh and Erlang(2) pattern. This data augmentation approach facilitates specification of the transitional measure in the Markov Chain. Gibbs steps are proposed to perform the Bayesian inference of such models. For model determination, we explored sum of relative error criterion that selects the best model. A numerical example with simulated data set is given.

• The Korean Journal of Applied Statistics
• /
• v.25 no.3
• /
• pp.415-422
• /
• 2012

For statistical microarray data analysis, clustering analysis is a useful exploratory technique and offers the promise of simultaneously studying the variation of many genes. However, most of the proposed clustering methods are not rigorously solved for a time-course microarray data cluster and for a fitting time covariate; therefore, a statistical method is needed to form a cluster and represent a linear trend of each cluster for each gene. In this research, we developed a modified hierarchical mixture of an experts model to suggest clustering data and characterize each cluster using a linear mixed effect model. The feasibility of the proposed method is illustrated by an application to the human fibroblast data suggested by Iyer et al. (1999).