Foyzul Hassan

Gender independent Bangla automatic speech recognition

Speaker-specific characteristics play an important role on the performance of Bangla (widely used as Bengali) automatic speech recognition (ASR). Gender factor shows adverse effect in the classifier while recognizing a speech by an opposite gender, such as, training a classifier by male but testing is done by female or vice-versa. To obtain a robust ASR system in practice it is necessary to invent a system that incorporates gender independent effect for particular gender. In this paper, we have proposed a Gender-Independent technique for ASR that focused on a gender factor. The proposed method trains the classifier with the both types of gender, male and female, and evaluates the classifier for the male and female. For the experiments, we have designed a medium size Bangla (widely known as Bengali) speech corpus for both the male and female. The proposed system has showed a significant improvement of word correct rates, word accuracies and sentence correct rates in comparison with the method that suffers from gender effects. Moreover, it requires a fewer mixture component in hidden Markov model (HMMs) and hence, computation time.

Bangla triphone HMM based word recognition

In this paper, we have prepared a medium size Bangla speech corpus and compare performances of different acoustic features for Bangla word recognition. Most of the Bangla automatic speech recognition (ASR) system uses a small number of speakers, but 40 speakers selected from a wide area of Bangladesh, where Bangla is used as a native language, are involved here. In the experiments, mel-frequency cepstral coefficients (MFCCs) are inputted to the triphone hidden Markov model (HMM) based classifiers for obtaining word recognition performance. From the experiments, it is shown that MFCC-based method of 39 dimensions provides a higher word correct rate (WCR) and word accuracy (WA) than the other methods investigated. Moreover, a higher WCR and WA is obtained by the MFCC39-based method with fewer mixture components in the HMM.

Local Feature or Mel Frequency Cepstral Coefficients - Which One Is Better for MLN-Based Bangla Speech Recognition?

This paper discusses the dominancy of local features (LFs), as input to the multilayer neural network (MLN), extracted from a Bangla input speech over mel frequency cepstral coefficients (MFCCs). Here, LF-based method comprises three stages: (i) LF extraction from input speech, (ii) phoneme probabilities extraction using MLN from LF and (iii) the hidden Markov model (HMM) based classifier to obtain more accurate phoneme strings. In the experiments on Bangla speech corpus prepared by us, it is observed that the LFbased automatic speech recognition (ASR) system provides higher phoneme correct rate than the MFCC-based system. Moreover, the proposed system requires fewer mixture components in the HMMs.

Gender Effects Suppression in Bangla ASR by Designing Multiple HMM-Based Classifiers

Speaker-specific characteristics play an important role on the performance of Bangla (widely used as Bengali) automatic speech recognition (ASR). It is difficult to recognize speech affected by gender factors, especially when an ASR system contains only a single acoustic model. If there exists any suppression process that represses the decrease of differences in acoustic-likelihood among categories resulted from gender factors, a robust ASR system can be realized. In this paper, we have proposed a technique of gender effects suppression that composed of two hidden Markov model (HMM)-based classifiers and that focused on a gender factor. In an experiment on Bangla speech database prepared by us, the proposed system has provided a significant improvement of word correct rate, word accuracy and sentence correct rate in comparison with the method that incorporates only a single HMM-based classifier for both male and female speakers.

Bangla ASR design by suppressing gender factor with gender-independent and gender-based HMM classifiers

Hidden factor such as gender characteristic plays an important role on the performance of Bangla (widely used as Bengali) automatic speech recognition (ASR). If there is a suppression process that represses the decrease of differences in acoustic-likelihood among categories resulted from gender factors, a robust ASR system can be realized. In our previous paper, we proposed a technique of gender effects suppression that composed of two hidden Markov model (HMM)-based classifiers that focused on a gender factor. In the proposed study, we have designed a new ASR for Bangla by suppressing the gender effects, which embeds three HMM-based classifiers for corresponding male, female and geneder-independent (GI) characteristics. In an experiment on Bangla speech database prepared by us, the proposed system that incorporates GI-classifier has achieved a significant improvement of word correct rate, word accuracy and sentence correct rate in comparison with our previous method that did not incorporate GI-classifier.

Bangla phoneme recognition using hybrid features

This paper presents a Bangla phoneme recognition method for Automatic Speech Recognition (ASR). The method consists of three stages: i) a multilayer neural network (MLN), which converts acoustic features, mel frequency cepstral coefficients (MFCCs), into phoneme probabilities, ii) the phoneme probabilities obtained from the first stage and corresponding Δ and ΔΔ are inserted into another MLN to improve the phoneme probabilities by reducing the context effect and (iii) the phoneme probabilities of current frame and corresponding MFCCs are fed into a hidden Markov model (HMM) based classifier to obtain more accurate phoneme strings. From the experiments on Bangla speech corpus prepared by us, it is observed that the proposed method provides higher phoneme recognition performance than the existing method. Moreover, it requires a fewer mixture components in the HMMs.

Phonetic Features enhancement for Bangla automatic speech recognition

This paper discusses a phonetic feature (PF) based automatic speech recognition system (ASR) for Bangla (widely known as Bengali), where the PF features are enhanced. There are three stages in this method where the first step maps Acoustic Features (AFs) or Local Features (LFs) into Phonetic Features (PFs) and the second step incorporates inhibition/enhancement (In/En) algorithm to change the PF dynamic patterns where patterns are enhanced for convex patterns and inhibited for concave patterns. The final step is for normalizing the extended PF vector using Gram-Schmidt algorithm and then passing through a Hidden Markov Model (HMM) based classifier. In our experiment on speech corpus for Bangla, the proposed feature extraction method provides higher sentence correct rate (SCR), word correct rate (WCR) and word accuracy (WA) compared to the methods that not incorporated In/En network.

Recurrent Neural Network Based Phoneme Recognition Incorporating Articulatory Dynamic Parameters

This paper describes a recurrent neural network (RNN) based phoneme recognition method incorporating articulatory dynamic parameters (Δ and ΔΔ). The method comprises three stages: (i) DPFs extraction using a recurrent neural network (RNN) from acoustic features, (ii) incorporation of dynamic parameters into a multilayer neural network (MLN) for reducing DPF context, and (iii) addition of an Inhibition/Enhancement (In/En) network for categorizing the DPF movement more accurately and Gram-Schmidt orthogonalization procedure for decorrelating the inhibited/enhanced data vector before connecting with a hidden Markov models (HMMs)-based classifier. From the experiments on Japanese Newspaper Article Sentences (JNAS), it is observed that the proposed method provides a higher phoneme correct rate over the method that does not incorporate dynamic articulatory parameters. Moreover, it reduces mixture components in HMM for obtaining a higher performance.

Japanese phonetic feature extraction for automatic speech recognition

This paper presents a method for extracting distinctive phonetic features (DPFs) for automatic speech recognition (ASR). The method comprises three stages: i) a acoustic feature extractor, ii) a multilayer neural network (MLN) and iii) a hidden Markov model (HMM) based classifier. At first stage, acoustic features, local features (LFs), are extracted from input speech. On the other stage, MLN generates a 45-dimentional DPF vector from the LFs of 75- dimentions. Finally, these 45-dimentional DPF vector is inserted into an HMM-based classifier to obtain phoneme strings. From the experiments on Japanese Newspaper Article Sentences (JNAS), it is observed that the proposed DPF extractor provides a higher phoneme correct rate and accuracy with fewer mixture components in the HMMs compared to the method based on mel frequency cepstral coefficients (MFCCs). Moreover, a higher correct rate for each phonetic feature is obtained using the proposed method.

Development of Analysis Rules for Bangla Part of Speech for Universal Networking Language

The Universal Networking Language (UNL) is a worldwide generalizes form human interactive in machine independent digital platform for defining, recapitulating, amending, storing and dissipating knowledge or information among people of different affiliations. The theoretical and practical research associated with these interdisciplinary endeavor facilities in a number of practical applications in most domains of human activities such as creating globalization trends of market or geopolitical independence among nations. In our research work we have tried to develop analysis rules for Bangla part of speech which will help to create a doorway for converting the Bangla language to UNL and vice versa and overcome the barrier between Bangla to other Languages.