Continuous Malayalam speech recognition using Hidden Markov Models

The authors recently designed and implemented a large-vocabulary, speaker-independent, continuous speech recognition system. The system is based on hidden Markov modeling (HMM) of phoneme-sized acoustic units using continuous mixture Gaussian densities. The main structure of the system is outlined with a focus on a method of generating mixture Gaussian density models through a merging procedure whose efficiency was recently improved significantly. The system has been evaluated on the TIMIT database on a task of vocabulary size 853 and various grammar perplexities. The word accuracies are 92.2%, 84.9%, and 60.1% for the test set perplexities of 25, 106, and 853 (no grammar), respectively. >

This paper describes an automatic text‐ and speaker‐independent language identification method based on hidden Markov models (HMMs) for acoustic features. The hidden Markov modeling is used to represent the phonotactics for each language. Each language has its proper phonotactics.The HMM topology here is a fully structured (ergodic) model wherein any state could transit to all states. Two kinds of HMMs are used: the discrete HMM (DHMM) with the codebook and the continuous density HMM (CHMM). The HMM was trained using both the Baum‐Welch (forward‐backward) algorithm and the Viterbi algorithm. The latter was used for emphasizing the state transition probability.For comparison, experiments also were conducted on the identification using a mixtured Gaussian distribution model with one state. This single‐state Gaussian distribution model gave the same performance as the HMM trained with the Baum‐Welch algorithm. This is because the transition between states which reflects the characteristics of each language does not affect the likelihood scores very much. This problem was addressed by emphasizing the transition probabilities and using the Viterbi algorithm, which resulted in an improvement in the recognition rates. The trigram for optimal state sequence is introduced. Combining it with the HMM produced the best results.

The main objective of this paper is to explore the effectiveness of perceptual features for performing isolated digits and continuous speech recognition. The proposed perceptual features are captured and code book indices are extracted. Expectation maximization algorithm is used to generate HMM models for the speeches. Speech recognition system is evaluated on clean test speeches and the experimental results reveal the performance of the proposed algorithm in recognizing isolated digits and continuous speeches based on maximum log likelihood value between test features and HMM models for each speech. Performance of these features is tested on speeches randomly chosen from “TI Digits_1”, “TI Digits_2” and “TIMIT” databases. This algorithm is tested for VQ and combination of VQ and HMM speech modeling techniques. Perceptual linear predictive cepstrum yields the accuracy of 86% and 93% for speaker independent isolated digit recognition using VQ and combination of VQ & HMM speech models respectively. This feature also gives 99% and 100% accuracy for speaker independent continuous speech recognition by using VQ and the combination of VQ & HMM speech modeling techniques.

Automatic speech recognition (ASR) has been extensively studied during the past few decades. Today, most of the ASR system based on statistical modelling, and HMM is the most popular one among them. Now days, performance of ASR become one of the major bottleneck for its practical use. Literature of ASR shows that optimal performance could be achieved by using Gaussian mixture hidden Markov model (HMM) but choice of Gaussian mixture is arbitrary with little justification. If we use the different number of Gaussian Mixture then we get different results. In this paper, we compare the performance of continuous Hindi speech recognition by using different number of Gaussian mixture. The aim of this paper is to investigate the optimal number of Gaussian mixture that exhibits maximum accuracy in the context of Hindi speech recognition. HMM toolkit HTK 3.4.1 is used for the implementation of this system, in which Mel frequency cepstral coefficient (MFCC) is used as a feature extraction technique. The experimental results show that the maximum performance of the proposed system is achieved when we use four component Gaussian mixtures HMM model.

An investigation into the use of Bayesian learning of the parameters of a multivariate Gaussian mixture density has been carried out. In a continuous density hidden Markov model (CDHMM) framework, Bayesian learning serves as a unified approach for parameter smoothing, speaker adaptation, speaker clustering, and corrective training. The goal of this study is to enhance model robustness in a CDHMM-based speech recognition system so as to improve performance. Our approach is to use Bayesian learning to incorporate prior knowledge into the CDHMM training process in the form of prior densities of the HMM parameters. The theoretical basis for this procedure is presented and preliminary results applying to HMM parameter smoothing, speaker adaptation, and speaker clustering are given.Performance improvements were observed on tests using the DARPA RM task. For speaker adaptation, under a supervised learning mode with 2 minutes of speaker-specific training data, a 31% reduction in word error rate was obtained compared to speaker-independent results. Using Baysesian learning for HMM parameter smoothing and sex-dependent modeling, a 21% error reduction was observed on the FEB91 test.

We present a hybrid speech recognition system for speaker independent continuous speech recognition. The system combines a novel information theory based neural network (NN) paradigm and discrete Hidden Markov models (HMMs) including state-of-the-art techniques like state clustered triphones. The novel NN type is trained by an algorithm based on principles of self-organization that achieves maximum mutual information between the generated output labels and the basic phonetic classes. The structure of the hybrid system is quite similar to a classical VQ-HMM system but the vector quantizer (VQ) is replaced by the NN. To evaluate the system we use the speaker independent part of the resource management (RM) database. We obtained an important improvement by introducing a novel kind of context dependent basic classes used by the acoustic processor. The average RM recognition result with a word-pair grammar is now 95.2% what is significantly better than a classical VQ-system, slightly better than a different hybrid system with a recurrent network as probability estimator, and very close to the best continuous probability density function (PDF) HMM speech recognizers.

An MLP/HMM hybrid model using nonlinear predictors

In this paper, we present time-sequence binary classifiers (TSBC) for phonetic feature extraction. A large neural network is divided into an array of TSBCs, and each TSBC is a multilayer neural network which is dedicatedly trained to extract low level acoustic features of only one phoneme category, resulting in lower neural network complexity. TSBC has a feature that its output units are sequentially arranged and trained to reflect the temporal information of phonetic features, which is very important in speech recognition. In our speaker-independent all-Chinese-Syllable continuous speech recognition system, TSBCs are efficiently combined with HMM techniques, where TSBCs are used to extract low level phonetic features and HMMs are used to recognize high level speech units. The evaluation experiments obtain 97.0% word accuracy for speaker-independent large-vocabulary and continuous speech recognition.

Continuous speech is far more natural and efficient than isolated speech for communication. However, for current state-of-the-art automatic speech recognition systems, isolated speech recognition (ISR) is far more accurate than continuous speech recognition (CSR). It is common practice in the speech research community to build CSR systems using only CS data. However, slowing of the speaking rate is a natural reaction for a user faced with the high error rates of current CSR systems. Ironically, CSR systems typically have a much higher word error rate when speakers slow down since the acoustic models are usually derived exclusively from continuous speech corpora. In this paper, we summarize our efforts to improve the robustness of our speaker-independent CSR system against speaking styles, without suffering a recognition accuracy penalty. In particular the multi-style trained system described in this paper attains a 7.0% word error rate for a test set consisting of both isolated and continuous speech, in contrast to the 10.9% word error rate achieved by the same system trained only on continuous speech.

The current state-of-the-art in large-vocabulary, continuous speech recognition is based on the use of hidden Markov models (HMM). In an attempt to improve over HMM performance, the authors developed a hybrid system that combines the advantages of neural networks and HMM using a multiple hypothesis (or N-best) paradigm. The connectionist component of the system, the segmental neural net (SNN), models all the frames of a phonetic segment simultaneously, thus overcoming the well-known conditional-independence limitation of the HMM. They describe the hybrid system and discuss various aspects of SNN modeling, including network architectures, training algorithms and context modeling. Finally, they evaluate the hybrid system by performing several speaker-independent experiments with the DARPA Resource Management (RM) corpus, and demonstrate that the hybrid system shows a consistent improvement in performance over the baseline HMM system. >

Hybrid methods which combine hidden Markov models (HMMs) and connectionist techniques take advantage of what are believed to be the strong points of each of the two approaches: the powerful discrimination-based learning of connectionist networks and the time-alignment capability of HMMs. Connectionist Viterbi Training (CVT) is a simple variation of Viterbi training which uses a back-propagation network to represent the output distributions associated with the transitions in the HMM. The work reported here represents the culmination of three years of investigation of various means by which HMMs and neural networks (NNs) can be combined for continuous speech recognition. This paper describes the CVT procedure, discusses the factors most important to its design and reports its recognition performance. Several changes made to the system over the past year are reported here, including: (1) the change from recurrent to non-recurrent NNs, (2) the change from Sphinx-style phone-based HMMs to word-based HMMS, (3) the addition of a corrective training procedure, and (3) the addition of an alternate model for every word. The CVT system, incorporating these changes, achieves 99.1% word accuracy and 98.0% string accuracy on the TI/NBS Connected Digits task (“TI Digits”).

In continuous speech recognition systems acoustical sub-units are used to represent word models. For this reason, phonetic transcription of training data and vocabulary words play a fundamental role in recognizers overall performance. In this work we investigate the influence of phonetic transcription effects both in the training data and in the vocabulary words in speaker independent continuous speech recognition systems.

Inherent features of the Bangla (widely used as Bengali) language like long and short vowels and many instances of allophones make it difficult to build a continuous speech recognizer for the language. Stress and accent vary in spoken Bangla language from region to region. But in formal read Bangla speech, stress and accents are ignored. There are three approaches to continuous speech recognition (CSR) based on the sub-word unit viz. word, phoneme and syllable. Pronunciation of words and sentences are strictly governed by set of linguistic rules. Many attempts have been made to build continuous speech recognizers for Bangla for small and restricted tasks. However, medium and large vocabulary CSR for Bangla is relatively new and not explored. In this paper, the authors have attempted for extracting local features (LFs) from a Bangla input speech for tri-phone based automatic speech recognition (ASR) method. The method comprises two stages, where the first stage extracts LFs from input speech and the final stage generates word strings based on trip hone hidden Markov models (HMMs). The objective of this research is to build a medium vocabulary trip hone based continuous speech recognizer for Bangla language by extracting LFs over mel frequency cepstral coefficients (MFCCs). In this experimentation using Bangla speech corpus prepared by us, the recognizer provides higher word accuracy, word correct rate as well as sentence correct rate for trained and tested sentences with fewer mixture components in HMMs.

In the electrocardiogram (ECG), the detection of QRS complexes is a fundamental step in the ECG signal processing chain since it allows the determination of other characteristics waves of the ECG and provides information about heart rate variability. In this work, an automatic QRS complex detector based on continuous density hidden Markov models (HMM) is proposed. HMM were trained using univariate observation sequences taken either from QRS complexes or their derivatives. The detection approach is based on the log-likelihood comparison of the observation sequence with a fixed threshold. A sliding window was used to obtain the observation sequence to be evaluated by the model. The threshold was optimized by receiver operating characteristic curves. Sensitivity (Sen), specificity (Spc) and F1 score were used to evaluate the detection performance. The approach was validated using ECG recordings from the MIT-BIH Arrhythmia database. A 6-fold cross-validation shows that the best detection performance was achieved with 2 states HMM trained with QRS complexes sequences (Sen = 0.668, Spc = 0.360 and F1 = 0.309). We concluded that these univariate sequences provide enough information to characterize the QRS complex dynamics from HMM. Future works are directed to the use of multivariate observations to increase the detection performance.

This paper aims to design and implement English digits speech recognition system using Matlab (GUI). This work was based on the Hidden Markov Model (HMM), which provides a highly reliable way for recognizing speech. The system is able to recognize the speech waveform by translating the speech waveform into a set of feature vectors using Mel Frequency Cepstral Coefficients (MFCC) technique This paper focuses on all English digits from (Zero through Nine), which is based on isolated words structure. Two modules were developed, namely the isolated words speech recognition and the continuous speech recognition. Both modules were tested in both clean and noisy environments and showed a successful recognition rates. In clean environment and isolated words speech recognition module, the multi-speaker mode achieved 99.5% whereas the speaker-independent mode achieved 79.5%. In clean environment and continuous speech recognition module, the multi-speaker mode achieved 72.5% whereas the speaker-independent mode achieved 56.25%. However in noisy environment and isolated words speech recognition module, the multi-speaker mode achieved 88% whereas the speaker-independent mode achieved 67%. In noisy environment and continuous speech recognition module, the multi-speaker mode achieved 82.5% whereas the speaker-independent mode achieved 76.67%. These recognition rates are relatively successful if compared to similar systems.