Chorkin Chan

Bayesian adaptive learning of the parameters of hidden Markov model for speech recognition

A theoretical framework for Bayesian adaptive training of the parameters of a discrete hidden Markov model (DHMM) and of a semi-continuous HMM (SCHMM) with Gaussian mixture state observation densities is presented. In addition to formulating the forward-backward MAP (maximum a posteriori) and the segmental MAP algorithms for estimating the above HMM parameters, a computationally efficient segmental quasi-Bayes algorithm for estimating the state-specific mixture coefficients in SCHMM is developed. For estimating the parameters of the prior densities, a new empirical Bayes method based on the moment estimates is also proposed. The MAP algorithms and the prior parameter specification are directly applicable to training speaker adaptive HMMs. Practical issues related to the use of the proposed techniques for HMM-based speaker adaptation are studied. The proposed MAP algorithms are shown to be effective especially in the cases in which the training or adaptation data are limited. >

Chinese word segmentation based on maximum matching and word binding force

A Chinese word segmentation algorithm based on forward maximum matching and word binding force is proposed in this paper. This algorithm plays a key role in post-processing the output of a character or speech recognizer in determining the proper word sequence corresponding to an input line of character images or a speech waveform. To support this algorithm, a text corpus of over 63 millions characters is employed to enrich an 80,000-words lexicon in terms of its word entries and word binding forces. As it stands now, given an input line of text, the word segmentor can process on the average 210,000 characters per second when running on an IBM RISC System/6000 3BT workstation with a correct word identification rate of 99.74%.

Isolated word recognition by neural network models with cross-correlation coefficients for speech dynamics

This paper presents an artificial neural network (ANN) for speaker-independent isolated word speech recognition. The network consists of three subnets in concatenation. The static information within one frame of speech signal is processed in the probabilistic mapping subnet that converts an input vector of acoustic features into a probability vector whose components are estimated probabilities of the feature vector belonging to the phonetic classes that constitute the words in the vocabulary. The dynamics capturing subnet computes the first-order cross correlation between the components of the probability vectors to serve as the discriminative feature derived from the interframe temporal information of the speech signal. These dynamic features are passed for decision-making to the classification subnet, which is a multilayer perceptron (MLP). The architecture of these three subnets are described, and the associated adaptive learning algorithms are derived. The recognition results for a subset of the DARPA TIMIT speech database are reported. The correct recognition rate of the proposed ANN system is 95.5%, whereas that of the best of continuous hidden Markov model (HMM)-based systems is only 91.0%. >

Postprocessing statistical language models for handwritten Chinese character recognizer

Two statistical language models have been investigated on their effectiveness in upgrading the accuracy of a Chinese character recognizer. The baseline model is one of lexical analytic nature which segments a sequence of character images according to the maximum matching of words with consideration of word binding forces. A model of bigram statistics of word-classes is then investigated and compared against the baseline model in terms of recognition rate improvement on the image recognizer. On the average, the baseline language model improves the recognition rate by about 7% while the bigram statistics model upgrades it by about 10%

A discrete contextual stochastic model for the off-line recognition of handwritten Chinese characters

We study a discrete contextual stochastic (CS) model for complex and variant patterns like handwritten Chinese characters. Three fundamental problems of using CS models for character recognition are discussed, and several practical techniques for solving these problems are investigated. A formulation for discriminative training of CS model parameters is also introduced and its practical usage investigated. To illustrate the characteristics of the various algorithms, comparative experiments are performed on a recognition task with a vocabulary consisting of 50 pairs of highly similar handwritten Chinese characters. The experimental results confirm the effectiveness of the discriminative training for improving recognition performance.

On-line adaptation of the SCHMM parameters based on the segmental quasi-Bayes learning for speech recognition

On-line quasi-Bayes adaptation of the mixture coefficients and mean vectors in semicontinuous hidden Markov model (SCHMM) is studied. The viability of the proposed algorithm is confirmed and the related practical issues are addressed in a specific application of on-line speaker adaptation using a 26-word English alphabet vocabulary.

Contextual vector quantization for speech recognition with discrete hidden Markov model

Abstract By using formulation of the finite mixture distribution identification, in this paper, several alternatives to the conventional LBG VQ method are investigated. A contextual VQ method based on the Markov Random Field (MRF) theory is proposed to model the speech feature vector space. Its superiority is confirmed by a series of comparative experiments in a speaker independent isolated word recognition task by using different VQ schemes as the front-end of DHMM. The motivation to use MRF to model the contextual dependence information in the underlying speech production process can be readily extended to acoustic modeling of the basic speech units in speech recognition.

The gradient projection method for the training of hidden Markov models

Abstract In this paper, the training of HMMs has been considered a general optimization problem with linear constraints. A gradient projection method for nonlinear programming with linear constraints has been introduced and presented to solve for “optimal” values of the model parameters. When this classic method is applied to train HMMs of discrete or Gaussian mixture observation densities, a very simple formulation can be derived due to the special structure of the constraints on the HMM parameters. This kind of classical gradient-based optimization methods can offer an opportunity for more flexible modeling of speech signals and more sophisticated training of model parameters for speech recognition.

Contextual vector quantization modeling of hand-printed Chinese character recognition

A hand-printed Chinese character recognizer based on contextual vector quantization (CVQ) has been built. The idea of CVQ is to quantize each pixel to a codeword by considering not just the pixel itself but its neighbors and their codeword identities as well. 100 samples of each character are collected from 100 writers, among them, 92 are used for training and 8 for testing. The characters are scanned by a 300 dpi scanner, which are then noise removed, thinned, segmented and size normalized. Stroke counts and segment strengths are adopted as observation features. For a vocabulary of 470 simplified Chinese characters, a recognition rate of 97% is achieved.

Separation of fricatives from aspirated plosives by means of temporal spectral variation

This paper discusses the separation of fricatives from aspirated plosives based on the dynamic properties of the consonant spectra. Twelve features of each CV syllable were constructed and employed to determine the discriminants for the two manners of articulation. Correct classifications were around 85 percent if the cases to be classified belonged to a set of CV syllables of 1) the same place of articulation under the same vowel environment; 2) the same vowel environment; or 3) the same place of articulation without using any information from the vowel region at all.