The present capabilities of speech recognition algorithms will be surveyed. The application of IC technology to the implementation of these algorithms will be explored and potential future directions will be determined.

Speech recognition

Detecting epileptic seizure from EEG signals constitutes a challenging problem of significant importance. Combining adaptive optimal kernel time-frequency representation and visibility graph, we develop a novel method for detecting epileptic seizure from EEG signals. We construct complex networks from EEG signals recorded from healthy subjects and epilepsy patients. Then we employ clustering coefficient, clustering coefficient entropy and average degree to characterize the topological structure of the networks generated from different brain states. In addition, we combine energy deviation and network measures to recognize healthy subjects and epilepsy patients, and further distinguish brain states during seizure free interval and epileptic seizures. Three different experiments are designed to evaluate the performance of our method. The results suggest that our method allows a high-accurate classification of epileptiform EEG signals.

Visibility Graph from Adaptive Optimal Kernel Time-Frequency Representation for Classification of Epileptiform EEG.

Diagnosis of schizophrenia (SZ) is traditionally performed through patient’s interviews by a skilled psychiatrist. This process is time-consuming, burdensome, subject to error and bias. Hence the aim of this study is to develop an automatic SZ identification scheme using electroencephalogram (EEG) signals that can eradicate the aforementioned problems and support clinicians and researchers. This study introduces a methodology design involving empirical mode decomposition (EMD) technique for diagnosis of SZ from EEG signals to perfectly handle the behavior of non-stationary and nonlinear EEG signals. In this study, each EEG signal is decomposed into intrinsic mode functions (IMFs) by the EMD algorithm and then twenty-two statistical characteristics/features are calculated from these IMFs. Among them, five features are selected as significant feature applying Kruskal Wallis test. The performance of the obtained feature set is tested through several renowned classifierson a SZ EEG dataset. Among the considered classifiers, theensemble bagged tree performed as the best classifier producing 93.21% correct classification rate for SZ, with an overall accuracy of 89.59% for IMF 2. These results indicate that EEG signals discriminate SZ patients from healthy control (HC) subjects efficiently and have the potential to become a tool for the psychiatrist to support the positive diagnosis of SZ.

A Computerized Method for Automatic Detection of Schizophrenia Using EEG Signals

Classification of imbalanced ECG beats using re-sampling techniques and AdaBoost ensemble classifier

Over the past decades, extensive research has been carried out on various possible implementations of automatic speech recognition (ASR) systems. The most renowned algorithms in the field of ASR are the mel-frequency cepstral coefficients and the hidden Markov models. However, there are also other methods, such as wavelet-based transforms, artificial neural networks and support vector machines, which are becoming more popular. This review article presents a comparative study on different approaches that were proposed for the task of ASR, and which are widely used nowadays.

https://ietresearch.onlinelibrary.wiley.com/doi/epdf/10.1049/iet-spr.2012.0151

Comparative study of automatic speech recognition techniques

Application of empirical mode decomposition and artificial neural network for the classification of normal and epileptic EEG signals

This paper presents a study of four different classifiers in the task of automatic speech based gender identification. Gender identification could have several applications in automatic speech and speaker recognition systems and in content -based multimedia indexing. Gaussian mixture model (GMM), multilayer perceptrons (MLP), vector quantization (VQ) and learning vector quantization (LVQ) are the classifiers used in this work along with mel frequency cepstral coefficients (MFCC). The performance attained by our best system is 96.4% identification accuracy using only 1s of speech per speaker using the IViE corpus.

A speech signal based gender identification system using four classifiers

As acoustic signal generated from vocal folds is directly af fected by vocal tract pathologies, it can be an effective too l f r diagnosis purpose. In this work, we present an efficient meth od for voice pathology detection based on speech signal proc essing and machine learning techniques. In the proposed method, we use d MFCC to represent the signal features, and we chose to combi ne GMM and SVM classifiers to benefit from their generative and discr iminative natures respectively. That is to exploit the simi larity function of the RBF kernel to separate the GMM models representing norma l and pathological voices. To further improve the separatio n, we used modified versions of the well known Kullback-leibler and Bha ttacharyya distances. The modified distances, unlike the cl assical ones, do satisfy all metric axioms. As a result, we obtained an impr ovement of 2 % and 4 % in terms of sensitivity compared to using the classical Kullback-leibler and Bhattacharyya distances r espectively. The Receiver Operating Curve (ROC) does illus trate the efficiency of the proposed method.

An Improved GMM-SVM System based on Distance Metric for Voice Pathology Detection

In this paper we present experiments we perform in order to recognize Arabic isolated words. Our recognition system is based on the combination of the supervised classifier at the classical system recognition based Markov modeling. This work is an alternative hybrid approach GHMM/supervised classifier (SVM or KNN) used in speech recognition using hidden Markov model with supervised classifier algorithm. The new approach GHMM/SVM or GHMM/KNN is introduced, evaluated and compared with traditional approach GHMM for isolated word recognition system. Both these approaches apply the same principles of feature extraction and time-sequence modeling; the principal difference lies in the additional classifier in the architecture used for training and recognition phases

New Hybrid System (Supervised Classifier/HMM) for Isolated Arabic Speech Recognition

Robust Speech Recognition Using Perceptual Wavelet Denoising and Mel-frequency Product Spectrum Cepstral Coefficient Features

/pdf/robust-speech-recognition-using-perceptual-wavelet-denoising-1v73yhyp9k.pdf

Hocine Bourouba

Papers

Application of empirical mode decomposition and artificial neural network for the classification of normal and epileptic EEG signals

A speech signal based gender identification system using four classifiers

An Improved GMM-SVM System based on Distance Metric for Voice Pathology Detection

New Hybrid System (Supervised Classifier/HMM) for Isolated Arabic Speech Recognition

Robust Speech Recognition Using Perceptual Wavelet Denoising and Mel-frequency Product Spectrum Cepstral Coefficient Features