Feature extraction and classification techniques for speaker recognition: A review

This paper investigates fault diagnosis in batch processes and presents a comparative study of feature extraction and classification techniques applied to a specific biotechnological case study: the fermentation process model by Birol et al. (Comput Chem Eng 26:1553-1565, 2002), which is a benchmark for advanced batch processes monitoring, diagnosis and control. Fault diagnosis is achieved using four approaches on four different process scenarios based on the different levels of noise so as to evaluate their effects on the performance. Each approach combines a feature extraction method, either multi-way principal component analysis (MPCA) or multi-way independent component analysis (MICA), with a classification method, either artificial neural network (ANN) or support vector machines (SVM). The performance obtained by the different approaches is assessed and discussed for a set of simulated faults under different scenarios. One of the faults (a loss in mixing power) could not be detected due to the minimal effect of mixing on the simulated data. The remaining faults could be easily diagnosed and the subsequent discussion provides practical insight into the selection and use of the available techniques to specific applications. Irrespective of the classification algorithm, MPCA renders better results than MICA, hence the diagnosis performance proves to be more sensitive to the selection of the feature extraction technique.

Speaker recognition is a biometric technique which uses individual speakers voice samples as a input for recognition purpose. Main goal of this work is to obtain better accuracy for speaker recognition system for large number of voice database. In this paper, a comparative study is made between various combinations of features for speaker identification system with feedforward artificial neural network(FFANN) and support vector machine for 10 and 20 speakers. Linear predictive coding, Mel frequency cepstral coefficient and zero crossing rate are used as a feature extraction techniques. Each features are tested separately and in combination with FFANN and SVM classifier on Matlab software. For ANN classifier 70% of total database are used for training,15% for validation and remaining 15% data are used for testing, number of hidden layers and number of neurons used are 2 and 80 respectively. It is observed from the result that efficiency of the system is not dropping by increasing number of speakers from 10(320 voice samples) to 20(640 voice samples) when using combination of LPC, MFCC, ZCR features with ANN classifier.

Chapter 10 - Parameterization techniques for automatic speech recognition system

Isolated Speech recognition systems (ISRS) have been implemented using microprocessors, digital signal processors and FPGAs and have been reported in the literature. In this paper, the study and implementation of an ISRS using Cypress Programmable System on Chip (PSoC) is presented. For the implementation, PSoC5 containing the ARM Cortex-M3 CPU is used. Recognition performance is studied using three feature extraction techniques (Zero crossing, Zero crossing with end point detection and Zero crossing with end point detection and Cochlear filter) and minimum distance classifier. The ISRS is trained and tested using 10 utterances of three words. The performance is evaluated for utterances by both single speaker and multiple speaker. From this study, it is found that the system using the Zero crossing with end point detection and Cochlear filter gives the highest recognition accuracy of 80%and 70% respectively for the utterances by single speaker and multiple speakers. Higher recognition accuracy can be obtained using more robust feature extraction and classification techniques at the cost of increase in the computation complexity and storage space. The proposed ISRS has applications in voice operated robots, appliances and access control systems.

Static and dynamic analysis are widely used in malware research. Although both approaches present considerable advantages according to their specific applications, they also present some challenges in terms of the time and resources constraints when extracting features. To alleviate those constraints, recent research works explore the use of techniques such as computer vision, in detecting malicious samples and their corresponding malware family. In this research, we aim to conduct a thorough comparison between a combination set of different feature extraction and classification techniques in order to identify the optimal approach that can be used in a real-time malware detection application portable executable (PE). Multiple Machine Learning (ML) and Deep Learning (DL) algorithms were used with image-based features and non image-based features. The results obtained show high performance when using static PE header features with K-Nearest Neighbor (KNN) and Bagged Trees algorithms for bi-level classification and Image-based Fast Fourier Transform (FFT) features with AlexNet for multi-level classification.

The rapid emergence of technology and big data science opened up a significant amount of work that has been carried out in the field of feature extraction and classification techniques of heart sound using various deep learning methods. Practically, medical practitioners use the same old scientific method and practice to seek out any cardiac disorders and predict any abnormality in the human heart. Heart sound normalization, denoising, segmentation, feature extraction, and classification techniques provide a suitable way of study of phonocardiography (PCG) signal analysis which eventually reduces the cost, makes the system compact, and simultaneously, can work with extensive training data. This paper mainly indulges in two parts feature extraction and classification. The proposed deep learning study for PCG signal used online available heart disease datasets, and time domain features like average energy, power, root mean square (RMS), total harmonic distortion, and zero Crossing rates are used. Statistical features used are kurtosis and skewness. The acoustic features used are Mel-frequency cepstrum coefficients (MFCCs), mel, chroma, contrast, and tonnetz. For the classification of heart sound, the proposed modified CNN-based mobile network is used. The modified CNN-based mobile network is very effective in heart sound analysis as it requires very less computation time and storage. The proposed CNN-based modified Mobile Network model attained an accuracy of 99.04 + 0.07% on the test dataset with a sensitivity of 96.8 + 0.03 % and specificity of 97.2 + 0.09%.

The aim of the study was to inspect the acoustic properties and sleep characteristics of a preapneic snoring sound. The feasibility of forecasting upcoming respiratory events by snoring sound was also investigated. Participants with habitual snoring or a heavy breathing sound during sleep were recruited consecutively. Polysomnography was conducted, and snoring-related breathing sound was recorded simultaneously. Acoustic features and sleep features were extracted from 30-second samples, and a machine learning algorithm was used to establish 2 prediction models. A total of 74 eligible participants were included. Model 1, tested by 5-fold cross-validation, achieved an accuracy of 0.92 and an area under the curve of 0.94 for respiratory event prediction. Model 2, with acoustic features and sleep information tested by Leave-One-Out cross-validation, had an accuracy of 0.78 and an area under the curve of 0.80. Sleep position was found to be the most important among all sleep features contributing to the performance of the 2 models. Preapneic sound presented unique acoustic characteristics, and snoring-related breathing sound could be deployed as a real-time apneic event predictor. The models, combined with sleep information, serve as a promising tool for an early warning system to forecast apneic events. Wang B, Yi X, Gao J, etal. Real-time prediction of upcoming respiratory events via machine learning using snoring sound signal. J Clin Sleep Med. 2021;17(9):1777-1784.

This paper examines the performance improvement via combining feature vectors for speaker verification. As feature vectors, we use LPC (linear predictive coding) spectral envelope (LPCSE), Mel-frequency cepstral coefficients (MFCC) and piecewise LPC (pLPC) pole distribution (pLPCPD). Here, pLPCPD is the feature which we have introduced and developed, and this paper shows a relationship of the underlying pLPC speech signal model for speaker verification with ARMA and AR models for speech production. We execute experiments of speaker verification with the combinations of the features. We examine the experimental results and show that (1) the smallest AER (average error rate) is achieved by MFCC followed in order by pLPCPD and LPCSE, (2) the AER achieved by LPCSE, MFCC and LPCSE+MFCC (combination of LPCSE and MFCC) is reduced by combining pLPCPD, (3) there are conflicting features in LPCSE, MFCC and pLPCPD to increase the AER, and the smallest AER is obtained by MFCC+pLPCPD among the combinations of LPCSE, MFCC and pLPCPD. We also investigate the relationship between pLPCPD and the aperiodicity used for vocal cord excitation in recent vocoders, which seems to explain the above result (2).

peech is the utmost communication medium for human beings which conveys rich and valuable information such as accent, gender, emotion and unique identity. Therefore, automatic speaker recognition can be developed based on unique characteristics of one’s speech and utilized for applications such as voice dialing, online banking, and telephone shopping to verify the identity of its users. However, retrieving salient features which are capable of identifying speakers is a challenging problem in speech recognition systems since speech contains abundant information. In this study, a total of 438 audio data obtained from speakers uttering speech in text-independent context is proposed using speech data elicited from three Malay male speakers. The performance of two popularly used feature extraction techniques namely, linear prediction coefficients (LPC) and Mel-frequency cepstral coefficients (MFCC) were compared using discriminant analysis model. Although both features yielded impressive outcomes, the MFCC features surpassed that of LPC by achieving a higher accuracy rate of 99.09%, which was 4.34% higher than the latter.

Speech signals are one of the most important means of communication among the human beings. In this paper, a comparative study of two feature extraction techniques are carried out for recognizing speaker independent spoken isolated words. First one is a hybrid approach with Linear Predictive Coding (LPC) and Artificial Neural Networks (ANN) and the second method uses a combination of Wavelet Packet Decomposition (WPD) and Artificial Neural Networks. Voice signals are sampled directly from the microphone and then they are processed using these two techniques for extracting the features. Words from Malayalam, one of the four major Dravidian languages of southern India are chosen for recognition. Training, testing and pattern recognition are performed using Artificial Neural Networks. Back propagation method is used to train the ANN. The proposed method is implemented for 50 speakers uttering 20 isolated words each. Both the methods produce good recognition accuracy. But Wavelet Packet Decomposition is found to be more suitable for recognizing speech because of its multi-resolution characteristics and efficient time frequency localizations.

Identifying a person's identity still often uses an ID card (KTP, SIM, passport, etc.). This method has a weakness because the ID Card is easily damaged and lost. Biometric recognition systems provide a solution by using human body parts as identity recognition. Sounds are readily available biometric information. Voice pattern recognition is used for the speaker identification process to obtain the identity of someone speaking. This paper reviews several feature extraction and classification methods that are often used in speaker identification. The selection of feature extraction methods and classification functions in computation and the level of accuracy of the speaker identification system. Based on the survey dataset applied with the feature extraction method, the Mel Frequency Cepstral Coefficients (MFCC) method has high accuracy even with noise input. Then in classification, the Gaussian Mixture Model (GMM) method is most often used because it can work in noise. Recently, a hybrid classifier has been developed, which increases the accuracy value.

This paper presents a robust speaker identification method from degraded speech signals. This proposed method depends on the Mel-frequency cepstral coefficients (MFCCs) for feature extraction from the degraded speech and its discrete cosine transform (DCT). It is known that the MFCCs based speech recognition methods are not robust enough in the presence of noise and channel degradation. So, the feature extraction from the DCT of the signal will assist in achieving a higher recognition rate. The artificial neural network (ANN) classification technique is used in the proposed method. The comparison between the proposed method and the method using the MFCCs only for feature extraction from noisy speech signals and telephone-like degraded signals shows that the proposed method improves the recognition rate in the presence of noise or degradation.

Speech recognition is gaining an increasing research interest in the last five decades. Speech processing is considered as an interdisciplinary branch of electronics and computer science domain. It considers speech as an input and converts it into the corresponding text. This paper describes the design and development of Marathi Numeric Speech Dataset. Marathi Numbers (Ank) ranging from Shunya(0) to Nau(9) and are taken into consideration for recording. Speech samples are collected from 50 native and 50 non-native speakers of Marathi language. The dataset remains as a gender balanced since it is recorded from 50 females and 50 male speakers. The age of speakers will affect the speech. Therefore, 5 different age groups such as 11-20, 21-30, 31-40, 41-50 and 51-60 are considered. Native and non-native speakers are selected to obtain ample amount of variations in the pronunciation of Marathi numerals. Feature extraction and feature matching technique plays a vital role for speech recognition and here LPC (Linear Predictive Coding) is used for extracting features from samples, whereas ANN (Artificial neural network) is used to classify them. Experimental specifications and results are also discussed. This research work has attempted to design and develop a speech recognition system, which can understand Marathi Ank (Numbers) and identify them accurately.

This paper explore on emotion recognition from Marathi speech signals by using feature extraction techniques and classifier to classify Marathi speech utterances according to their emotional contains. A different type of speech feature vectors contains different emotions, due to their corresponding natures. In this we have categorized the emotions as namely Anger, Happy, Sad, Fear, Neutral and Surprise. Mel Frequency Cepstral Coefficient (MFCC) feature parameters extracted from Marathi speech Signals depend on speaker, spoken word as well as emotion. Gaussian mixture Models (GMM) is used to develop Emotion classification model. In this, recently proposed feature extraction technique and classifier is used for Marathi spoken words. In this each subject/Speaker has spoken 7 Marathi words with 6 different emotions that 7 Marathi words are Aathawan, Aayusha, Chamakdar, Iishara, Manav, Namaskar, and Uupay. For experimental work we have created total 924 Marathi speech utterances database and from this we achieved the empirical performance of overall emotion recognition accuracy rate obtained using MFCC and GMM is 84.61% rate of our Emotion Recognition for Marathi Spoken Words (ERFMSW) system. We got average accuracy for male and female is 86.20% and 83.03% respectively.KeywordsEmotion RecognitionMel Frequency Cepstral CoefficientGaussian mixture modelsspeaking rateMarathi Speech Emotional Database

Lately, 3D imaging techniques have achieved a lot of progress due to recent developments in 3D sensor technologies. This leads to a great interest regarding 3D image feature extraction and classification techniques. As pointed out in literature, one of the most important and discriminative features in images is the textural content. Within this context, we propose a texture feature extraction technique for volumetric images with improved discrimination power. The method could be used in textured volumetric data classification tasks. To achieve this, we fuse two complementary pieces of information, feature vectors derived from Local Binary Patterns (LBP) and the Gray-Level Co-occurrence Matrix-based methods. They provide information regarding the image pattern and the contrast, homogeneity and local anisotropy in the volumetric data, respectively. The performance of the proposed technique was evaluated on a public dataset consisting of volumetric textured images affected by several transformations. The classifiers used are the Support Vector Machine, k-Nearest Neighbours and Random Forest. Our method outperforms other handcrafted 3D or 2D texture feature extraction methods and typical deep-learning networks. The proposed technique improves the discrimination power and achieves promising results even if the number of images per class is relatively small.