Abstract. This paper presents a stochastic segmental speech recognizer that models the a posteriori probabilities directly. The main issues concerning the system are segmental phoneme classification, utterance-level aggregation and the pruning of the search space. For phoneme classification artificial neural networks and support vector machines are applied. Phonemic segmentation and utterance-l...

The phoneme classification inaccuracy at the acoustic phonetic level is a major weakness in most speech recognition systems. However, the inaccuracy will violate phonotactic constraints at the acoustic phonetic level. A better performance is expected if a language model is adopted in a recognition system for post-processing phoneme estimates and making corrections with a set of explicit rules o...

The most popular neural network models for use in speech recognition experiments are employ model neurons which apply a nonlinear function to a weighted sum of their inputs. These networks are trained by adjusting the weights in the weighted sums. There is another class of models called Boolean networks, in which the model neurons output logical functions of their inputs. The training process a...

Improving phoneme recognition has attracted the attention of many researchers due to its applications in various fields of speech processing. Recent research achievements show that using deep neural network (DNN) in speech recognition systems significantly improves the performance of these systems. There are two phases in DNN-based phoneme recognition systems including training and testing. Mos...

the relation between the variants of the frequency and the length of the word is among the laws that george kinsley zipf discovered, according to which, the longer is the length of a word, the less is the frequency of its application in the text. it is tried in the present research to examine this law, which has not been so far examined on the arabic language of the qur’ān, with a concentration...

This paper investigates a number of ensemble methods for improving the performance of phoneme classification for use in a speech recognition system. Two ensemble methods are described; boosting and mixtures of experts, both in isolation and in combination. Results are presented on two speech recognition databases: an isolated word database and a large vocabulary continuous speech database. Thes...

In this paper we introduce two ideas for phoneme classification: First, we derive the necessary steps to integrate linear transform into the computation of reproducing kernels. This concept not restricted to phoneme classification and can be applied in a wider range of research subjects. Second, in the context of support vector machine (SVM) classification, correlation features based on MFCC-ve...

K-nearest neighbor (k-NN) classification is a powerful and simple method for classification. k-NN classifiers approximate a Bayesian classifier for a large number of data samples. The accuracy of k-NN classifier relies on the distance metric used for calculating nearest neighbor and features used for instances in training and testing data. In this paper we use deep neural networks (DNNs) as a f...

Phoneme classification is investigated for linear feature domains with the aim of improving robustness to additive noise. In linear feature domains noise adaptation is exact, potentially leading to more accurate classification than representations involving non-linear processing and dimensionality reduction. A generative framework is developed for isolated phoneme classification using linear fe...