Variable resolution spectral/temporal features for automatic speech recognition.

Significance of analytic phase of speech signals in speaker verification

In this paper, we present an efficient parallel implementation of Mel-frequency Cepstral Coefficient (MFCC)-based feature extraction and describe the optimizations required for effective throughput on Graphics Processing Units (GPU) processors. We demonstrate that the feature extraction process in automatic speech recognition is well suited for GPUs and a substantial reduction in computation time can be obtained by performing feature extraction on these platforms. Using a single Nvidia GTX580 GPU our proposed approach is shown to be approximately 90x faster than a sequential CPU implementation, enabling feature extraction to be performed at under 0.01% real-time. This is significantly faster than prior reported results implemented on GPUs, DSPs and FPGAs. Furthermore we demonstrate that multiple MFCC features can be generated for a set of predefined Vocal-Tract-Length-Normalization (VTLN) alpha parameters with little degradation in throughput. Using the approach described in this paper MFCC features were extracted in 0.05% and 0.09% realtime, for 11 and 21 VTLN parameters respectively. (4 pages)

In this paper, we propose a feature set by concatenating Mel-Frequency Cepstral Coefficients (MFCC) and Spectral Centroid Magnitude Coefficients (SCMC) features for Acoustic Scene Classification (ASC) using Deep Neural Networks (DNN). MFCC features are used to hold the acoustic characteristics such as spectral envelope of an acoustic scene in each frame. It also carries the sub-band average energy as a single dimension. SCMC features are used to hold the distribution of energy in a sub-band effectively. A test is carried out on Tampere University of Technology (TUT) Acoustic Scenes 2017 Dataset. The DNN architecture for utterance level classification has been used. The proposed system’s performance on a 4-fold cross-validation setup is 80.2% and it gives 5.4% relative improvement in performance when compared to the baseline system that uses log-Mel band energies with Multi-Layer Perceptron model.

Feature extraction techniques for input representation to diagnose infant diseases have received significant attention recently. Mel Frequency Cepstral Coefficients (MFCC) is one of the most popular feature extraction techniques due to its representation method being very similar to the human auditory system. The MFCC method for feature extraction depends on several important parameter settings, namely the number of filter banks, and the number of coefficients used in the final representation. These settings affects the way the features are represented, and in turn, affects the performance of the classifier for diagnosis of the disease. In this paper, the Particle Swarm Optimization (PSO) algorithm was used to optimise the parameters of the MFCC feature extraction method for classifying infants with asphyxia. The extracted MFCC features were then used to train several MLP classifiers over different initialization values. The accuracy of these classifiers was then used to guide the PSO optimization. Our results show that the optimization of MFCC computation using PSO yielded 93.9% accuracy, an improvement of 1.45% over typical MFCC parameter settings using the same classifier.

This paper presents the use of auditory perception based admissible wavelet packet tree (WPT) for partitioning of speech frequencies into different bands based on the Mel scale or the Bark Scale. The proposed WPTs selected using root mean square error (RMSE) criterion mimic the Mel scale or the bark scale more accurately and hence the human auditory system. Performance of the features obtained from the proposed WPTs is compared with Mel frequency cepstral coefficients (MFCC). The algorithms are evaluated using NIST TI-46 isolated-word database using hidden Markov model (HMM) as a classifier. Experimental results show that the performance of proposed features is better than MFCC and other wavelet features for isolated word recognition (IWR).

Speaker identification refers to the process of recognizing human voice using artificial intelligence techniques. Speaker identification technologies are widely applied in voice authentication, security and surveillance, electronic voice eavesdropping, and identity verification. In the speaker identification process, extracting discriminative and salient features from speaker utterances is an important task to accurately identify speakers. Various features for speaker identification have been recently proposed by researchers. Most studies on speaker identification have utilized short-time features, such as perceptual linear predictive (PLP) coefficients and Mel frequency cepstral coefficients (MFCC), due to their capability to capture the repetitive nature and efficiency of signals. Various studies have shown the effectiveness of MFCC features in correctly identifying speakers. However, the performances of these features degrade on complex speech datasets, and therefore, these features fail to accurately identify speaker characteristics. To address this problem, this study proposes a novel fusion of MFCC and time-based features (MFCCT), which combines the effectiveness of MFCC and time-domain features to improve the accuracy of text-independent speaker identification (SI) systems. The extracted MFCCT features were fed as input to a deep neural network (DNN) to construct the speaker identification model. Results showed that the proposed MFCCT features coupled with DNN outperformed existing baseline MFCC and time-domain features on the LibriSpeech dataset. In addition, DNN obtained better classification results compared with five machine learning algorithms that were recently utilized in speaker recognition. Moreover, this study evaluated the effectiveness of one-level and two-level classification methods for speaker identification. The experimental results showed that two-level classification presented better results than one-level classification. The proposed features and classification model for identifying a speaker can be widely applied to different types of speaker datasets.

In the field of medicine, with the introduction of computer systems that can collect and analyze massive amounts of data, many non-invasive diagnostic methods are being developed for a variety of conditions. In this study, our aim is to develop a non-invasive method of classifying respiratory sounds that are recorded by an electronic stethoscope and the audio recording software that uses various machine learning algorithms.In order to store respiratory sounds on a computer, we developed a cost-effective and easy-to-use electronic stethoscope that can be used with any device. Using this device, we recorded 17,930 lung sounds from 1630 subjects.We employed two types of machine learning algorithms; mel frequency cepstral coefficient (MFCC) features in a support vector machine (SVM) and spectrogram images in the convolutional neural network (CNN). Since using MFCC features with a SVM algorithm is a generally accepted classification method for audio, we utilized its results to benchmark the CNN algorithm. We prepared four data sets for each CNN and SVM algorithm to classify respiratory audio: (1) healthy versus pathological classification; (2) rale, rhonchus, and normal sound classification; (3) singular respiratory sound type classification; and (4) audio type classification with all sound types. Accuracy results of the experiments were; (1) CNN 86%, SVM 86%, (2) CNN 76%, SVM 75%, (3) CNN 80%, SVM 80%, and (4) CNN 62%, SVM 62%, respectively.As a result, we found out that spectrogram image classification with CNN algorithm works as well as the SVM algorithm, and given the large amount of data, CNN and SVM machine learning algorithms can accurately classify and pre-diagnose respiratory audio.

This paper describes a study of a set of features based on cepstral mean and variance normalization (CMVN) plus auto regressive moving average (ARMA) filtering technique which is called MVA and on histogram equalization (HEQ) for robust speech recognition. First, we use MVA then HEQ in combination with CMVN and ARMA filtering as a post-processing module to mel frequency cepstral coefficients (MFCC), Relative Spectral-Perceptual linear prediction (RASTA-PLP) and power normalized cepstral coefficients (PNCC) features to improve the performance of the automatic speech recognition (ASR) system. The results on the Arabic database task have shown that both methods MVA and HEQ+ARMA improves the success rate for all features compared to the baseline system however HEQ was not found to perform better than MVA. The results also provide that RASTA-PLP outperforms PNCC and MFCC features.

This work investigates the significance of the voiced and unvoiced region for detecting common cold from the speech signal. In literature, the entire speech signal is processed to detect the common cold and other diseases. This study uses a short-time energy-based approach to segment the voiced and unvoiced region of the speech signal. Then, frame-wise mel frequency cepstral coefficients (MFCC) features are extracted from the voiced and unvoiced segments of each speech utterance, and statistics (mean, variance, skewness, and kurtosis) are calculated to get the feature vector for each speech utterance. The support vector machine (SVM) is utilized to analyze the performance of features extracted from the voiced and unvoiced region. Result shows that the feature extracted from voiced segments, unvoiced segments, and complete active speech (CAS) gives almost similar results using the MFCC features and SVM classifier. Therefore, rather than processing the CAS, we can process the unvoiced speech segments, which have fewer frames compared to CAS and voiced regions of speech. The processing of solely unvoiced segments can reduce the time and computation complexity of a speech signal-based common cold detection system.

We present a Conformer-based end-to-end neural diarization (EEND) model that uses both acoustic input and features derived from an automatic speech recognition (ASR) model. Two categories of features are explored: features derived directly from ASR output (phones, position-in-word and word boundaries) and features derived from a lexical speaker change detection model, trained by fine-tuning a pretrained BERT model on the ASR output. Three modifications to the Conformer-based EEND architecture are proposed to incorporate the features. First, ASR features are concatenated with acoustic features. Second, we propose a new attention mechanism called contextualized self-attention that utilizes ASR features to build robust speaker representations. Finally, multi-task learning is used to train the model to minimize classification loss for the ASR features along with diarization loss. Experiments on the two-speaker English conversations of Switchboard+SRE data sets show that multi-task learning with position-in-word information is the most effective way of utilizing ASR features, reducing the diarization error rate (DER) by 20% relative to the baseline.

Speech and language recognition is one of the most important requirements in daily life. Most people in the world use English as the highly preferable language for easy communication. As a result, English is the highest speaking language in the World. Speech is the easiest way of medium for communication. Hunan-human, human-machine, and machine-machine communication are very much dependent on extensive and robust speech recognition. There are various important features of speech. The accent is one of them. It is a unique feature of speech that varies according to region and dialect. Speech scientists are continuously working on the improvement of speech recognition techniques and performance. Various challenges are being faced day by day and new areas of research are opened to overcome those difficulties. Variation of accent is one of the biggest challenges in improving speech recognition performances. In this work, the native and non-native English accent has been considered. The classification task has been performed on the English speech of non-native speakers and native American English speakers. Three sets of male and female speaker data both for native and non-native speech have been collected from the Voice Cloning Toolkit corpus of the Centre for Speech Technology Research. Mel Frequency Cepstral Coefficient features were extracted from the audio signals and given as input parameters to the supervised machine learning models. Evaluation metrics have been generated, multiple results were produced and compared.

This paper addresses the development of an Automatic Speech Recognition (ASR) system for the Moroccan Dialect. Dialectal Arabic (DA) refers to the day-to-day vernaculars spoken in the Arab world. In fact, Moroccan Dialect is very different from the Modern Standard Arabic (MSA) because it is highly influenced by the French Language. It is observed throughout all Arab countries that standard Arabic widely written and used for official speech, news papers, public administration and school but not used in everyday conversation and dialect is widely spoken in everyday life but almost never written. we propose to use the Mel Frequency Cepstral Coefficient (MFCC) features to specify the best speaker identification system. The extracted speech features are quantized to a number of centroids using vector quantization algorithm. These centroids constitute the codebook of that speaker. MFCC’s are calculated in training phase and again in testing phase. Speakers uttered same words once in a training session and once in a testing session later. The Euclidean distance between the MFCC’s of each speaker in training phase to the centroids of individual speaker in testing phase is measured and the speaker is identified according to the minimum Euclidean distance. The code is developed in the MATLAB environment and performs the identification satisfactorily.

The mel-frequency cepstral coefficient (MFCC) or perceptual linear prediction (PLP) feature extraction typically used for automatic speech recognition (ASR) employ several principles which have known counterparts in the cochlea and auditory nerve: frequency decomposition, mel- or bark-warping of the frequency axis, and compression of amplitudes. It seems natural to ask if one can profitably employ a counterpart of the next physiological processing step, synaptic adaptation. We, therefore, incorporated a simplified model of short-term adaptation into MFCC feature extraction. We evaluated the resulting ASR performance on the AURORA 2 and AURORA 3 tasks, in comparison to ordinary MFCCs, MFCCs processed by RASTA, and MFCCs processed by cepstral mean subtraction (CMS), and both in comparison to and in combination with Wiener filtering. The results suggest that our approach offers a simple, causal robustness strategy which is competitive with RASTA, CMS, and Wiener filtering and performs well in combination with Wiener filtering. Compared to the structurally related RASTA, our adaptation model provides superior performance on AURORA 2 and, if Wiener filtering is used prior to both approaches, on AURORA 3 as well.

This paper presents the automatic genre classification of Indian Tamil music and western music using timbral features and fractional Fourier transform (FrFT) based Mel frequency cepstral coefficient (MFCC) features. The classifier model for the proposed system has been built using K-nearest neighbours and support vector machine (SVM) classifiers. In this work, the performance of various features extracted from music excerpts have been analyzed, to identify the appropriate feature descriptors for the two major genres of Indian Tamil music, namely classical music (Carnatic based devotional hymn compositions) and folk music. The results have shown that the feature combination of spectral roll off, spectral flux, spectral skewness and spectral kurtosis, combined with fractional MFCC features, outperforms all other feature combinations, to yield a higher classification accuracy of 96.05 %, as compared to the accuracy of 84.21 % with conventional MFCC. It has also been observed, that the FrFT based MFCC, with timbral features and SVM, efficiently classifies the two western genres of rock and classical music, from the GTZAN dataset, with fewer features and a higher classification accuracy of 96.25 %, as compared to the classification accuracy of 80 % with conventional MFCC.

Conventional Speech Emotion Recognition (SER) approaches put more emphasis on extracting magnitude spectrum-based features, such as Mel Frequency Cepstral Coefficients (MFCCs), and Mel spectrogram. However, phase information is ignored due to signal processing difficulties such as the phase wrapping issue. This work develops a multichannel Convolution Neural Network-Bidirectional Long Short Term Memory (CNN-BLSTM) architectures with an attention mechanism for speaker-independent SER by considering phase and magnitude spectrum-based features. The phase-based features are extracted using the Modified Group Delay Function (MODGD). The obtained phase features are combined with MFCC features. The CNN-BLSTM network extract learned representation from magnitude and phase features. The learned representation from MFCCs and MODGD are combined and given as an input to the Support Vector Machine (SVM) for classification. The Deep Canonical Correlation Analysis (DCCA) is used to maximize the correlation between magnitude and phase information to improve the conventional SER system’s performance. Here the IEMOCAP database is used for performance analysis. The experimental results show improvement over MFCC features and existing approaches for unimodal SER. In this work, we also developed real-time Web server application for the proposed architecture.