• Journal of the Institute of Convergence Signal Processing
• /
• v.15 no.2
• /
• pp.37-41
• /
• 2014

The vector Taylor series (VTS) based method usually employs clean speech Hidden Markov Models (HMMs) when compensating speech feature vectors or adapting the parameters of trained HMMs. It is well-known that noisy speech HMMs trained by the Multi-condition TRaining (MTR) and the Multi-Model-based Speech Recognition framework (MMSR) method perform better than the clean speech HMM in noisy speech recognition. In this paper, we propose a method to use the noise-adapted HMMs in the VTS-based speech feature compensation method. We derived a novel mathematical relation between the train and the test noisy speech feature vector in the log-spectrum domain and the VTS is used to estimate the statistics of the test noisy speech. An iterative EM algorithm is used to estimate train noisy speech from the test noisy speech along with noise parameters. The proposed method was applied to the noise-adapted HMMs trained by the MTR and MMSR and could reduce the relative word error rate significantly in the noisy speech recognition experiments on the Aurora 2 database.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.4 no.2
• /
• pp.236-240
• /
• 2004

Stroke-based composite HMMs with articulation states are proposed to deal with 3D spatio-temporal trajectory gestures. The direct use of 3D data provides more naturalness in generating gestures, thereby avoiding some of the constraints usually imposed to prevent performance degradation when trajectory data are projected into a specific 2D plane. Also, the decomposition of gestures into more primitive strokes is quite attractive, since reversely concatenating stroke-based HMMs makes it possible to construct a new set of gesture HMMs without retraining their parameters. Any deterioration in performance arising from decomposition can be remedied by a partial tuning process for such composite HMMs.

• The Journal of the Acoustical Society of Korea
• /
• v.10 no.2
• /
• pp.29-35
• /
• 1991

In this paper, we discuss the method of recognizing allophone unit isolated words using hidden Markov model(HMM). Frist we constructed allophone lexicon by extracting allophones from training data and by training allophone HMMs. And then to recognize isolated words using allophone HMMs, it is necessary to construct word dictionary which contains information of allophone sequence and inter-allophone transition probability. Allophone sequences are represented by allophone HMMs. To see the effects of inter-allophone transition probability and to determine optimal probabilities, we performend some experiments. And we showed that small number of traing data and simple train procedure is needed to train word HMMs of allophone sequences and that not less performance than word unit HMM is obtained.

• The Journal of the Acoustical Society of Korea
• /
• v.15 no.4E
• /
• pp.21-27
• /
• 1996

In this paper, we propose a discriminative training algorithm for the stochastic segment model (SSM) in continuous speech recognition. As the SSM is usually trained by maximum likelihood estimation (MLE), a discriminative training algorithm is required to improve the recognition performance. Since the SSM does not assume the conditional independence of observation sequence as is done in hidden Markov models (HMMs), the search space for decoding an unknown input utterance is increased considerably. To reduce the computational complexity and starch space amount in an iterative training algorithm for discriminative SSMs, a hybrid architecture of SSMs and HMMs is programming using HMMs. Given the segment boundaries, the parameters of the SSM are discriminatively trained by the minimum error classification criterion based on a generalized probabilistic descent (GPD) method. With the discriminative training of the SSM, the word error rate is reduced by 17% compared with the MLE-trained SSM in speaker-independent continuous speech recognition.

• ETRI Journal
• /
• v.32 no.1
• /
• pp.160-162
• /
• 2010

For the acoustic models of embedded speech recognition systems, hidden Markov models (HMMs) are usually quantized and the original full space distributions are represented by combinations of a few quantized distribution prototypes. We propose a maximum likelihood objective function to train the quantized distribution prototypes. The experimental results show that the new training algorithm and the link structure adaptation scheme for the quantized HMMs reduce the word recognition error rate by 20.0%.

• Journal of KIISE:Software and Applications
• /
• v.30 no.9
• /
• pp.867-875
• /
• 2003

This paper concerns continuous density HMM topology optimization. There have been several researches for HMM topology optimization. BIC (Bayesian Information Criterion) is one of the well known optimization criteria, which assumes statistically well behaved homogeneous model parameters. HMMs, however, are composed of several different kind of parameters to accommodate complex topology, thus BIC's assumption does not hold true for HMMs. Even though BIC reduced the total number of parameters of HMMs, it could not improve the recognition rates. In this paper, we proposed two new model selection criteria, HBIC (HMM-oriented BIC) and BIC_Anti. The former is proposed to improve BIC by estimating model priors separately. The latter is to combine BIC and anti-likelihood to accelerate discrimination power of HMMs. We performed some comparative research on couple of model selection criteria for online handwriting data recognition. We got better recognition results with fewer number of parameters.

• Journal of KIISE:Software and Applications
• /
• v.36 no.9
• /
• pp.683-690
• /
• 2009

One of the major problems that prevent us from building a good recognition system for large-scale on-line Chinese character recognition using HMMs is increasing recognition time. In this paper, we propose a clustering method to solve recognition speed problem and an efficient distance measure between HMMs. From the experiments, we got about twice the recognition speed and 95.37% 10-candidate recognition accuracy, which is only 0.9% decrease, for 20,902 Chinese characters defined in Unicode CJK unified ideographs.

• Journal of Institute of Control, Robotics and Systems
• /
• v.17 no.8
• /
• pp.747-752
• /
• 2011

HMMs (Hidden Markov Models) are widely used for biological signal, such as EEG (electroencephalogram) sequence, analysis because of their ability to incorporate sequential information in their structure. A recent trends of research are going after the biological interpretable HMMs, and we need to control the complexity of the HMM so that it has good generalization performance. So, an automatic means of optimizing the structure of HMMs would be highly desirable. In this paper, we described a procedure of classification of motor imagery EEG signals using HMM. The motor imagery related EEG signals recorded from subjects performing left, right hand and foots motor imagery. And the proposed a method that was focus on the validation of the HSA (Harmony Search Algorithm) based optimization for HMM. Harmony search algorithm is sufficiently adaptable to allow incorporation of other techniques. A HMM training strategy using HSA is proposed, and it is tested on finding optimized structure for the pattern recognition of EEG sequence. The proposed HSA-HMM can performs global searching without initial parameter setting, local optima, and solution divergence.

• Journal of Electrical Engineering and Technology
• /
• v.7 no.1
• /
• pp.109-114
• /
• 2012

In this paper, we utilize training strategy of hidden Markov model (HMM) to use in versatile issues such as classification of time-series sequential data such as electric transient disturbance problem in power system. For this, an automatic means of optimizing HMMs would be highly desirable, but it raises important issues: model interpretation and complexity control. With this in mind, we explore the possibility of using genetic algorithm (GA) and harmony search (HS) algorithm for optimizing the HMM. GA is flexible to allow incorporating other methods, such as Baum-Welch, within their cycle. Furthermore, operators that alter the structure of HMMs can be designed to simple structures. HS algorithm with parameter-setting free technique is proper for optimizing the parameters of HMM. HS algorithm is flexible so as to allow the elimination of requiring tedious parameter assigning efforts. In this paper, a sequential data analysis simulation is illustrated, and the optimized-HMMs are evaluated. The optimized HMM was capable of classifying a sequential data set for testing compared with the normal HMM.

• The KIPS Transactions:PartB
• /
• v.14B no.7
• /
• pp.523-530
• /
• 2007

This paper proposes a new stochastic optimization algorithm for hidden Markov models (HMMs) used as a recognizer of automatic lipreading. The proposed method combines a global stochastic optimization method, the simulated annealing technique, and the local optimization method, which produces fast convergence and good solution quality. We mathematically show that the proposed algorithm converges to the global optimum. Experimental results show that training HMMs by the method yields better lipreading performance compared to the conventional training methods based on local optimization.