• The Journal of the Acoustical Society of Korea
• /
• v.23 no.1
• /
• pp.75-81
• /
• 2004

Gaussian selection (GS) is a popular approach in the continuous density hidden Markov model for fast decoding. It enables fast likelihood computation by reducing the number of Gaussian components calculated. In this paper, we propose a new GS method for the phonetic tied-mixture (PTM) hidden Markov models. The PTM model can represent each state of the same topological location with a shared set of Gaussian mixture components and contort dependent weights. Thus the proposed method imposes constraint on the weights as well as the number of Gaussian components to reduce the computational load. Experimental results show that the proposed method reduces the percentage of Gaussian computation to 16.41%, compared with 20-30% for the conventional GS methods, with little degradation in recognition.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2007.11a
• /
• pp.469-473
• /
• 2007

For the quantitative analysis of inclusion using OES data, separating of noise and inclusion is needed. In previous methods assuming that noises come from a normal distribution, intensity levels beyond a specific threshold are determined as inclusions. However, it is not possible to classify inclusions in low intensity region using this method, even though every inclusion is an element of some chemical compound. In this paper, we assume that distribution of OES data is a Gaussian mixture and estimate the parameters of the mixture model using EM algorithm. Then, we calculate mixing ratio of noise and inclusion using these parameters to separate noise and inclusion.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.9C
• /
• pp.853-858
• /
• 2006

TMHMM(Tied Mixture Hidden Markov Model) is an important approach to reduce the number of free parameters in speech recognition. However, this model suffers from a degradation in recognition accuracy due to its GPDF (Gaussian Probability Density Function) clustering error. This paper proposes a clustering algorithm, called HCNN(Homogeneous Centroid Neural network), to cluster acoustic feature vectors in TMHMM. Moreover, the HCNN uses the heterogeneous distance measure to allocate more code vectors in the heterogeneous areas where probability densities of different states overlap each other. When applied to Korean digit isolated word recognition, the HCNN reduces the error rate by 9.39% over CNN clustering, and 14.63% over the traditional K-means clustering.

• Journal of Korean Society of Transportation
• /
• v.28 no.3
• /
• pp.169-183
• /
• 2010

Image-based traffic information collection systems have entered widespread adoption and use in many countries since these systems are not only capable of replacing existing loop-based detectors which have limitations in management and administration, but are also capable of providing and managing a wide variety of traffic related information. In addition, these systems are expanding rapidly in terms of purpose and scope of use. Currently, the utilization of image processing technology in the field of traffic accident management is limited to installing surveillance cameras on locations where traffic accidents are expected to occur and digitalizing of recorded data. Accurately recording the sequence of situations around a traffic accident in a freeway and then objectively and clearly analyzing how such accident occurred is more urgent and important than anything else in resolving a traffic accident. Therefore, in this research, existing technologies, this freeway attribute, velocity changes, volume changes, occupancy changes reflect judge the primary. Furthermore, We pointed out by many past researches while presenting and implementing an active and environmentally adaptive methodology capable of effectively reducing false detection situations which frequently occur even with the Gaussian Mixture model analytical method which has been considered the best among well-known environmental obstacle reduction methods. Therefore, in this way, the accident was the final decision. Also, environmental factors occur frequently, and with the index finger situations, effectively reducing that can actively and environmentally adaptive techniques through accident final judgment. This implementation of the evaluate performance of the experiment road of 12 incidents in simulated and the jang-hang IC's real-time accident experiment. As a result, the do well detection 93.33%, false alarm 6.7% as showed high reliability.

• Journal of Communications and Networks
• /
• v.18 no.2
• /
• pp.182-189
• /
• 2016

The focus in this paper is on obtaining tight, simple algebraic-form bounds and invertible expressions for the average symbol error probability (ASEP) of M-ary phase shift keying (MPSK) in a class of composite fading channels. We employ the mixture gamma (MG) distribution to approximate the signal-to-noise ratio (SNR) distributions of fading models, which include Nakagami-m, Generalized-K ($K_G$), and Nakagami-lognormal fading as specific examples. Our approach involves using the tight upper and lower bounds that we recently derived on the Gaussian Q-function, which can easily be averaged over the general MG distribution. First, algebraic-form upper bounds are derived on the ASEP of MPSK for M > 2, based on the union upper bound on the symbol error probability (SEP) of MPSK in additive white Gaussian noise (AWGN) given by a single Gaussian Q-function. By comparison with the exact ASEP results obtained by numerical integration, we show that these upper bounds are extremely tight for all SNR values of practical interest. These bounds can be employed as accurate approximations that are invertible for high SNR. For the special case of binary phase shift keying (BPSK) (M = 2), where the exact SEP in the AWGN channel is given as one Gaussian Q-function, upper and lower bounds on the exact ASEP are obtained. The bounds can be made arbitrarily tight by adjusting the parameters in our Gaussian bounds. The average of the upper and lower bounds gives a very accurate approximation of the exact ASEP. Moreover, the arbitrarily accurate approximations for all three of the fading models we consider become invertible for reasonably high SNR.

• Proceedings of the KIEE Conference
• /
• 2006.10c
• /
• pp.470-472
• /
• 2006

Sensor networks are the results of convergence of very important technologies such as wireless communication and micro electromechanical systems. In recent years, sensor networks found a wide applicability in various fields such as environment and health, industry scene system monitoring, etc. A very important step for these many applications is pattern classification and recognition of data collected by sensors installed or deployed in different ways. But, pattern classification and recognition are sometimes difficult to perform. Systematic approach to pattern classification based on modem learning techniques like Multivariate Gaussian mixture models, can greatly simplify the process of developing and implementing real-time classification models. This paper proposes a new recognition system which is hierarchically composed of many sensor nodes having the capability of simple processing and wireless communication. The proposed system is able to perform context classification of sensed data using the Multivariate Gaussian function. In order to verify the usefulness of the proposed system, it was applied to intelligent dust collecting system.

• ETRI Journal
• /
• v.33 no.6
• /
• pp.949-952
• /
• 2011

In this letter, a simplified suboptimum receiver based on soft-limiting for the detection of binary antipodal signals in non-Gaussian noise modeled as a generalized normal-Laplace (GNL) distribution combined with Gaussian noise is presented. The suboptimum receiver has low computational complexity. Furthermore, when the number of diversity branches is small, its performance is very close to that of the Neyman-Pearson optimum receiver based on the probability density function obtained by the Fourier inversion of the characteristic function of the GNL-plus-Gaussian distribution.

• The Korean Journal of Applied Statistics
• /
• v.22 no.4
• /
• pp.759-770
• /
• 2009

A problem of separating signals from noises is considered, when they are randomly mixed in the observation. It is assumed that the noise follows a Gaussian distribution and the signal follows a Gamma distribution, thus the underlying distribution of an observation will be a mixture of Gaussian and Gamma distributions. The parameters of the mixture model will be estimated from the EM algorithm. Then the signals and noises will be classified by a fixed threshold approach based on multiple testing using positive false discovery rate and Bayes error. The proposed method is applied to a real optical emission spectroscopy data for the quantitative analysis of inclusions. A simulation is carried out to compare the performance with the existing method using 3 sigma rule.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.47 no.1
• /
• pp.22-29
• /
• 2010

A real-time detection and tracking of moving objects in video sequences is very important for smart surveillance systems. In this paper, we propose a novel algorithm for the detection of moving objects that is the entropy-based adaptive Gaussian mixture model (AGMM). First, the increment of entropy generally means the increment of complexity, and objects in unstable conditions cause higher entropy variations. Hence, if we apply these properties to the motion segmentation, pixels with large changes in entropy in moments have a higher chance in belonging to moving objects. Therefore, we apply the Clausius entropy theory to convert the pixel value in an image domain into the amount of energy change in an entropy domain. Second, we use an adaptive background subtraction method to detect moving objects. This models entropy variations from backgrounds as a mixture of Gaussians. Experiment results demonstrate that our method can detect motion object effectively and reliably.

• Journal of Korea Multimedia Society
• /
• v.15 no.5
• /
• pp.577-586
• /
• 2012

The smartphone emerged due to the rapid development of the Internet has brought greater convenience to life in a positive manner. Recently, however, because of unconstrained exposure to harmful video, reckless use of smart phones has become a domestic issue in our society. In this paper, a service which detects harmful videos by using the acoustic information is designed and implemented on the Android OS. In order to implement the service of Android OS-based detection of the harmful movie, the speed of existing sound-based detection method for harmful videos is improved. The GMM(Gaussian Mixture Model) was used for classifier and the number of Gaussian Mixture was 18. The implemented service shows a detection rate of 97.02% for a total of 1,210 data files (approximately 687 hours) which comprises 669 general videos files (about 424 hours) and 541 harmful video files (about 263 hours). It's speed is 5.6 times faster than the traditional methods whitout reducing the detection rate.