Albino Nogueiras

The demiphone: an efficient contextual subword unit for continuous speech recognition

In this paper, we introduce the demiphone as a context-dependent phonetic unit for continuous speech recognition. A phoneme is divided into two parts: a left demiphone that accounts for the left coarticulation and a right demiphone that copes with the right-hand side context. This unit discards the dependence between the effects of both side contexts, but it models the transition between phonemes as the triphone does. By concatenating a left demiphone and a right demiphone a triphone can be built, although the left and the right-context coarticulations are modeled independently. The main appeal of this unit stems from its reduced number (respect to the number of triphones) and its capability to model left and right contexts unseen together in the training material. Thus, the demiphone shares in a simple way the advantages of a smoothed parameter estimation with the ability of generalization. In the present work, the demiphone is motivated and experimentally supported. Furthermore, demiphones are compared with triphones smoothed and generalized by decision-tree state-tying, accepted as the most powerful tool for coarticulation modeling at the present state of the art. The main conclusion of our work is that the demiphone simplifies the recognition system and yields a better performance than the triphone, at least for small or moderate size databases. This result may be explained by the ability of the demiphone to provide an excellent trade-off between a detailed coarticulation modeling and a proper parameter estimation.

Duration modeling with expanded HMM applied to speech recognition

The occupancy of the HMM states is modeled by means of a Markov chain. A linear estimator is introduced to compute the probabilities of the Markov chain. The distribution function (DF) represents accurately the observed data. Representing the DF as a Markov chain allows the use of standard HMM recognizers. The increase of complexity is negligible in training and strongly limited during recognition. Experiments performed on acoustic-phonetic decoding shows how the phone recognition rate increases from 60.6 to 61.1. Furthermore, on a task of database inquires, where phones are used as subword units, the correct word rate increases from 88.2 to 88.4.

Explicit segmentation of speech using Gaussian models

The authors investigate an automatic method to segment labeled speech. The method needs an initial estimation of the segmentation which is provided by an alignment based on HMM. Afterwards, the boundaries are refined moving the frontier frames to the segment which is more similar to the speech frame. Gaussian PDFs are used as a similarity measure. The performance of the method is evaluated using the TIMIT database. If boundary deviations (from the reference position) larger than 20 ms are counted as errors, then the replacement of the boundaries reduces the error by 30%. Additional experiments show how the proposed method makes the performance independent of the speaker dependent or speaker independent data used to estimate the HMM.

Frequency and time filtering of filter-bank energies for HMM speech recognition

In speech recognition, a discriminative frequency weighting can be achieved by decorrelating the frequency sequence of log mel-scaled filter-bank energies with a computationally inexpensive filter. We show how the spectral parameters that result from this kind of frequency filtering, both alone and combined with filtering of their time trajectories, are competitive with respect to the conventional cepstral representations of speech signals.

Multidialectal Spanish acoustic modeling for speech recognition

During the last years, language resources for speech recognition have been collected for many languages and specifically, for global languages. One of the characteristics of global languages is their wide geographical dispersion, and consequently, their wide phonetic, lexical, and semantic dialectal variability. Even if the collected data is huge, it is difficult to represent dialectal variants accurately. This paper deals with multidialectal acoustic modeling for Spanish. The goal is to create a set of multidialectal acoustic models that represents the sounds of the Spanish language as spoken in Latin America and Spain. A comparative study of different methods for combining data between dialects is presented. The developed approaches are based on decision tree clustering algorithms. They differ on whether a multidialectal phone set is defined, and in the decision tree structure applied. Besides, a common overall phonetic transcription for all dialects is proposed. This transcription can be used in combination with all the proposed acoustic modeling approaches. Overall transcription combined with approaches based on defining a multidialectal phone set leads to a full dialect-independent recognizer, capable to recognize any dialect even with a total absence of training data from such dialect. Multidialectal systems are evaluated over data collected in five different countries: Spain, Colombia, Venezuela, Argentina and Mexico. The best results given by multidialectal systems show a relative improvement of 13% over the results obtained with monodialectal systems. Experiments with dialect-independent systems have been conducted to recognize speech from Chile, a dialect not seen in the training process. The recognition results obtained for this dialect are similar to the ones obtained for other dialects.

Multi-dialectal Spanish speech recognition

Spanish is a global language, spoken in a big number of different countries with a big dialectal variability‥ This paper deals with the suitability of using a single multi-dialectal acoustic modeling for all the Spanish variants spoken in Europe and Latin America. This paper deals with the suitability of using a single multi-dialectal acoustic modeling for all the Spanish variants spoken in Europe and Latin America. The objective is two fold. First, it allows to use all the available databases to jointly train and improve the same system. Second, it allows to use a single system for all the Spanish speakers. The paper describes the rule- based phonetic transcription used for each dialectal variant, the selection of the shared and the specific phonemes to be modeled in a multi-dialectal recognition system, and the results of a multi-dialectal system dealing with dialects in and out of the training set.