A standard model creating device for providing a high-precision standard model used for pattern recognition such as speech recognition, character recognition, or image recognition by using a probability model based on a hidden Markov model, the Bayesian theory, linear discrimination analysis, intention interpretation using a probability model such as a Bayesian net, and data-mining performed by using a probability model. The standard model creating device comprises a reference model preparing unit (102) for preparing one or more reference models, a reference model storage unit (103) storing a reference model (121) prepared by the reference model preparing unit (102), and a standard model creating unit (104) for creating a standard model (122) by calculating a statistic quantity of the standard model in such a way that the probability or likelihood of one or more reference models stored in the reference model storage unit (103) is maximized or locally maximized.

Standard model creating device and standard model creating method

Widespread use of large-vocabulary continuous speech recognition systems has recently occurred, encouraging the application of speech recognition techniques to various problems. One of the factors that adversely affect the performance of speech recognition systems is a mismatch between the acoustic properties of the speech of the system user and the acoustic model. The speech of young or middle-aged adults is generally used in constructing the acoustic model. Thus, a mismatch occurs between the model and the acoustic properties of the speech of the elderly, which may degrade the recognition rate. In this study, a large-scale elderly speech database (200 sentences ×301 subjects) is used to train the acoustic model, and the resulting elderly acoustic model is evaluated by using a large-vocabulary continuous speech recognition system. In the experiments, the word recognition rate was improved by 3 to 5% compared to the recognition results of an acoustic model trained by young or middle-aged adult speech, namely, by the JNAS speech database (150 sentences ×260 subjects, average 28.6 years). It is also verified experimentally that the recognition rate is further improved in speaker adaptation to elderly speech by making use of an acoustic model trained by elderly speech. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 2, 87(7): 49–57, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjb.20101

Acoustic models of the elderly for large‐vocabulary continuous speech recognition

A standard model creating apparatus which provides a high-precision standard model used for pattern recognition such as speech recognition, character recognition, or image recognition using a probability model based on a hidden Markov model, Bayesian theory, or linear discrimination analysis; intention interpretation using a probability model such as a Bayesian net; data-mining performed using a probability model; and so forth. The standard model creating apparatus includes a reference model preparing unit that prepares at least one reference model; a reference model storing unit that stores the reference model prepared by the reference model preparing unit; and a standard model creating unit that creates a standard model by calculating statistics of the standard model so as to maximize or locally maximize the probability or likelihood with respect to the reference model stored in the reference storing unit.

Standard-model generation for speech recognition using a reference model

Statistical speech recognition using continuousdensity hidden Markov models (CDHMMs) has yielded many practical applications. However, in general, mismatches between the training data and input data significantly degrade recognition accuracy. Various acoustic model adaptation techniques using a few input utterances have been employed to overcome this problem. In this article, we survey these adaptation techniques, including maximum a posteriori (MAP) estimation, maximum likelihood linear regression (MLLR), and eigenvoice.

Speaker adaptation techniques for automatic speech recognition

The automatic recognition of disordered speech is a domain that is characterised by limited amounts of training data for each speaker and large intra- and inter-speaker variations. This paper is concerned with how best to train an acoustic models in these circumstances; in particular, we look at how to select data for a background model from a pool of speakers for a given target speaker. We show that rather than including data from all available speakers (the standard approach in the typical speech domain), significantly better accuracy can be achieved by carefully selecting which speakers should contribute. Different methods based on measuring acoustic closeness between speakers and ranking them accordingly are investigated, and on the UASpeech isolated word recognition task, we achieve a 11.5% relative improvement compared to the baseline which uses data from all speakers. Accuracies for speakers with moderate to severe impairments are shown to improve the most with one speaker classed as having ‘low’ intelligibility gaining a 60% relative improvement in accuracy.

Automatic selection of speakers for improved acoustic modelling: recognition of disordered speech with sparse data

Describes an efficient method for unsupervised speaker adaptation. This method is based on (1) selecting a subset of speakers who are acoustically close to a test speaker, and (2) calculating adapted model parameters according to the previously stored sufficient HMM statistics of the selected speakers' data. In this method, only a few unsupervised test speaker's data are required for the adaptation. Also, by using the sufficient HMM statistics of the selected speakers' data, a quick adaptation can be done. Compared with a pre-clustering method, the proposed method can obtain a more optimal speaker cluster because the clustering result is determined according to test speaker's data on-line. Experimental results show that the proposed method attains better improvement than MLLR from the speaker independent model. Moreover the proposed method utilizes only one unsupervised sentence utterance, while MLLR usually utilizes more than ten supervised sentence utterances.

M. Yamada

Papers

Unsupervised speaker adaptation based on sufficient HMM statistics of selected speakers