Abstract not found

decide which contexts are similar and can share parameters. A key feature of this approach is that it allows the construction of models which are dependent upon contextual effects occurring across word boundaries. The use of cross word context dependent models presents problems for conventional decoders. The second part of the thesis therefore presents a new decoder design which is capable of using these models efficiently. The decoder is suitable for use with very large vocabularies and long span language models. It is also capable of generating a lattice of word hypotheses with little computational overhead. These lattices can be used to constrain further decoding, allowing efficient use of complex acoustic and language models. The effectiveness of these techniques has been assessed on a variety of large vocabulary continuous speech recognition tasks and results are presented which analyse performance in terms of computational complexity and recognition accuracy. The experiments dem

Abstract not found

The benefits of a speech recognition machine would be many, resulting in the improvement of the quality of life for people. The design of a speech recognition system can be divided into two parts, commonly known as the front-end and back-end. The front-end deals with the conversion of the analog speech signal into features for classification. This thesis investigates optimal feature-sets for speech recognition. The objectives for an optimal feature-set are improved recognition performance, noise robustness, talker insensitivity and efficiency. Three problems that make it difficult to find optimal features are: 1) the amount of resources (time and computations) required to evaluate the performance of a feature-set, 2) the size of the feature space, and 3) the dependence of features upon some words in t...

This study deals with acoustic phonetic modelling in HMM based continuous speech recognition. Context dependent phone models were derived by a decision tree clustering algorithm. In particular, lexical stress was introduced as a clustering variable in addition to the phonetic context. The parameter sharing model was extended by tying HMM states across different target phones. For instance, one or more states of a tense vowel and the corresponding lax vowel were tied if they proved to be acoustically similar. The results indicate that the use of lexical stress information in acoustic modelling might be fruitful when large amounts of training data are available. 1. INTRODUCTION State of the art speech recognition typically makes use of context dependent HMMs. These units are often triphones or even quinphones. However, to model all existing contexts for many speakers is an unfeasible task due to the enormous amounts of data such a task would require. Therefore, parameters are often sha...

Phone-like units (PLUs) for automatic speech recognition are derived using a decision tree algorithm. In our approach we use information such as target phone label, immediate context, lexical stress level and function word affiliation in the decision tree analysis. The resulting PLUs are shown to improve phone and word recognition.

Abstract not found

using lexical stress and context Högberg, J. and Sjölander, K.