• Journal of Digital Convergence
• /
• v.12 no.7
• /
• pp.273-278
• /
• 2014

In vocabulary recognition using HMM(Hidden Markov Model) by model for the observation of a discrete probability distribution indicates the advantages of low computational complexity, but relatively low recognition rate. Improve them with a HMM model is proposed for the optimization of the Bayesian methods. In this paper is posterior distribution and prior distribution in recognition Gaussian mixtures model provides a model to optimize of the Bayesian methods vocabulary recognition. The result of applying the proposed method, the recognition rate of 97.9% in vocabulary recognition, respectively.

• Journal of Korea Multimedia Society
• /
• v.3 no.6
• /
• pp.566-574
• /
• 2000

Although a numbed of variants of 2D HMM have been proposed in the literature, they are, in a word, too simple to model the variabilities of images for diverse classes of objects; they do not realize the modeling capability of the 1D HMM in 2D. Thus the author thinks they are poor substitutes for the HMM in 2D. The new model proposed in this paper is a hidden Markov lattice or, we can dare say, a 2D HMM with the causality of top-down and left-right direction. Then with the addition of a lattice constraint, the two algorithms for the evaluation of a model and the maximum likelihood estimation of model parameters are developed in the theoretical perspective. It is a more natural extension of the 1D HMM. The proposed method will provide a useful way of modeling highly variable patterns such as offline cursive characters.

• The Journal of the Acoustical Society of Korea
• /
• v.32 no.2
• /
• pp.174-180
• /
• 2013

This paper proposes a decision tree based context clustering algorithm for HMM-based speech synthesis systems using the cross likelihood ratio with a hierarchical prior (CLRHP). Conventional algorithms tie the context-dependent HMM states that have similar statistical characteristics, but they do not consider the statistical similarity of split child nodes, which does not guarantee the statistical difference between the final leaf nodes. The proposed CLRHP algorithm improves the reliability of model parameters by taking a criterion of minimizing the statistical similarity of split child nodes. Experimental results verify the superiority of the proposed approach to conventional ones.

• Journal of Korea Multimedia Society
• /
• v.6 no.1
• /
• pp.48-57
• /
• 2003

This paper describes an approach for classifying myoelectric patterns using a multilayer perceptrons (MLP's) with genetic algorithm and hidden Markov models (HMM's) hybrid classifier. Genetic Algorithms play a role of selecting Multilayer Perceptron's optimized initial connection weights by its typical global search. The dynamic aspects of EMG are important for tasks such as continuous prosthetic control or various time length EMG signal recognition, which have not been successfully mastered by the most neural approaches. It is known that the hidden Markov model (HMM) is suitable for modeling temporal patterns. In contrast, the multilayer feedforward networks are suitable for static patterns. And, a lot of investigators have shown that the HMM's to be an excellent tool for handling the dynamical problems. Considering these facts, we suggest the combination of ANN and HMM algorithms that might lead to further improved EMG recognition systems.

• The Journal of the Acoustical Society of Korea
• /
• v.18 no.8
• /
• pp.94-99
• /
• 1999

It is the hybrid structure of HMM and neural network(NN) that shows high recognition rate in speech recognition algorithms. And it is a method which has majorities of statistical model and neural network model respectively. In this study, we propose a new style of the hybrid structure of semi-continuous HMM(SCHMM) and radial basis function(RBF), which re-estimates weighting coefficients probability affecting observation probability after Baum-Welch estimation. The proposed method takes account of the similarity of basis Auction of RBF's hidden layer and SCHMM's probability density functions so as to discriminate speech signals sensibly through the learned and estimated weighting coefficients of RBF. As simulation results show that the recognition rates of the hybrid structure SCHMM/RBF are higher than those of SCHMM in unlearned speakers' recognition experiment, the proposed method has been proved to be one which has more sensible property in recognition than SCHMM.

• The KIPS Transactions:PartB
• /
• v.10B no.2
• /
• pp.229-236
• /
• 2003

Named entity recognition is the process indispensable to question answering and information extraction systems. This paper presents an HMM based named entity (m) recognition method using the construction principles of compound words. In Korean, many named entities can be decomposed into more than one word. Moreover, there are contextual relationships among nouns in an NE, and among an NE and its surrounding words. In this paper, we classify words into a word as an NE in itself, a word in an NE, and/or a word adjacent to an n, and train an HMM based on NE-related word types and parts of speech. Proposed named entity recognition (NER) system uses trigram model of HMM for considering variable length of NEs. However, the trigram model of HMM has a serious data sparseness problem. In order to solve the problem, we use multi-level back-offs. Experimental results show that our NER system can achieve an F-measure of 87.6% in the economic articles.

• Journal of Internet Computing and Services
• /
• v.10 no.4
• /
• pp.55-70
• /
• 2009

Vision-based Human computer interaction is an emerging field of science and industry to provide natural way to communicate with human and computer. In that sense, inferring the emotional state of the person based on the facial expression recognition is an important issue. In this paper, we present a novel approach to recognize facial expression from a sequence of input images using emotional specific HMM (Hidden Markov Model) and facial motion tracking based on optical flow. Conventionally, in the HMM which consists of basic emotional states, it is considered natural that transitions between emotions are imposed to pass through neutral state. However, in this work we propose an enhanced transition framework model which consists of transitions between each emotional state without passing through neutral state in addition to a traditional transition model. For the localization of facial features from video sequence we exploit template matching and optical flow. The facial feature displacements traced by the optical flow are used for input parameters to HMM for facial expression recognition. From the experiment, we can prove that the proposed framework can effectively recognize the facial expression in real time.

• Journal of KIISE:Computing Practices and Letters
• /
• v.11 no.3
• /
• pp.284-291
• /
• 2005

I propose a framework for detecting, recognizing and classifying facial features based on learned expression patterns. The framework recognizes facial expressions by using PCA and expression HMM(EHMM) which is Hidden Markov Model (HMM) approach to represent the spatial information and the temporal dynamics of the time varying visual expression patterns. Because the low level spatial feature extraction is fused with the temporal analysis, a unified spatio-temporal approach of HMM to common detection, tracking and classification problems is effective. The proposed recognition framework is accomplished by applying posterior probability between current visual observations and previous visual evidences. Consequently, the framework shows accurate and robust results of recognition on as well simple expressions as basic 6 facial feature patterns. The method allows us to perform a set of important tasks such as facial-expression recognition, HCI and key-frame extraction.

• The Journal of the Acoustical Society of Korea
• /
• v.34 no.2
• /
• pp.171-176
• /
• 2015

We present a speaker verification method by extracting i-vectors based on dominant state information of Hidden Markov Model (HMM) - Universal Background Model (UBM). Ergodic HMM is used for estimating UBM so that various characteristic of individual speaker can be effectively classified. Unlike Gaussian Mixture Model(GMM)-UBM based speaker verification system, the proposed system obtains i-vectors corresponding to each HMM state. Among them, the i-vector for feature is selected by extracting it from the specific state containing dominant state information. Relevant experiments are conducted for validating the proposed system performance using the National Institute of Standards and Technology (NIST) 2008 Speaker Recognition Evaluation (SRE) database. As a result, 12 % improvement is attained in terms of equal error rate.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.3
• /
• pp.273-278
• /
• 2004

In this paper, the recurrent neural networks (RNN) is applied to compensate for HMM recognition algorithm, which is commonly used as main recognizer. Among these recurrent neural networks, the multi-layer recurrent neural prediction model (MRNPM), which allows operating in real-time, is used to implement learning and recognition, and HMM and MRNPM are used to design a hybrid-type main recognizer. After testing the designed speech recognition algorithm with Korean number pronunciations (13 words), which are hardly distinct, for its speech-independent recognition ratio, about 5% improvement was obtained comparing with existing HMM recognizers. Based on this result, only optimal (recognition) codes were extracted in the actual DSP (TMS320C6711) environment, and the embedded speech recognition system was implemented. Similarly, the implementation result of the embedded system showed more improved recognition system implementation than existing solid HMM recognition systems.