Speaker Verification Using Line Spectrum Frequency, Formant, and Support Vector Machine

This paper proposes an automatic stream-weight and threshold estimation method for noise-robust speaker verification using multi-stream HMMs integrating segmental and prosodic information. The proposed method simultaneously optimizes stream-weights and a decision threshold by combining the linear discriminant analysis (LDA) and Adaboost techniques. Experiments were conducted using Japanese connected digit speech contaminated by white noise with various SNRs. In this experiment, a target ratio of false acceptance rate (FAR) and false rejection rate (FRR) was set by 1:1 so as to adjust them to approach an equal error rate (EER). Experimental results show that the proposed method effectively estimates stream-weights and thresholds so that FARs and FRRs are adjusted to EERs in most of the SNR conditions

This article explores a method for speaker verification to appropriately set the threshold value used to judge the identities of individual speakers and also considers methods to update speaker models and the robustness of the speaker models so as to make the models more robust with respect to utterance variations, using a small amount of data for updating that was recently uttered. The speaker model is represented by the hidden Markov model (HMM). In the model updating, the parameters of the speaker HMM are estimated from the data for updating and the current parameter values. For setting the threshold, the new threshold value for each speaker is set with the initial value as a value that is passed to an FA rate higher than the equal error rate (a value for which the false rejection rate and the false acceptance rate are equal), which is calculated from the data for updating and which steadily approaches the value that was passed to the equal error rate in concert with updating of the speaker HMM. The results of evaluating this method using text-dependent and text-prompted speaker verification experiments with twenty speakers shows that the average error rate fell by roughly 40% for the text-independent type and by roughly 80% for the text-prompted type as compared to when the model and threshold value are not updated. © 1999 Scripta Technica, Syst Comp Jpn, 30(13): 96–105, 1999

The problem of speaker recognition from a given set of speakers for any language and any context is considered. A database of Russian numerals that contains speech segments from 216 men and 177 women, each of whom spoke from 400 to 800 words, is used for recognition. Speech has been recorded on different types of microphones in different rooms at the natural noise level. Recognition is based on solutions of the inverse problem of finding the voice excitation pulse shape for each pitch period by the known speech segment. The pulse shape is defined as the inverse Fourier transform of the regularized ratio of speech signal spectra at the intervals of the open and closed glottis. Recognition is carried out by ten parameters: the pitch period, the open glottis interval duration, times when the source amplitude is maximum, minimum, or zero, the amplitude ratio for the minimum and maximum source pulses, three decomposition ratios of the source function by the principal component method, and the vowel duration. In such a recognition procedure, in the case of the utterance of a word that contains one vowel, the false reject rate (FRR) for men is 1.7---5.4%, and the false acceptance rate (FAR) is 5.4---7.1%. For women FRR = 2---5.2% and FAR = 5.2---6.3%. The recognition error decreases with an increasing number of vowels in the speech signal. At 10 vowels, for men FRR = 0.05---0.2% and FAR = 0.07---0.8%, and for women FRR = 0.09---0.2% and FAR = 0.17---2.1%.

This paper describes different approaches for the face authentication from the features and classification abilities point of view. Authors compare two types of features - Histogram of Oriented Gradients (HOG) and Local Binary Patterns (LBP) including their combination. These parameters are classified using Multilayer Neural Network (MLNN) and Support Vector Machines (SVM). Face authentication consists of several steps. The first step contains Viola-Jones algorithm for face detection. Authors resize the detected face for a fixed vector and afterwards, it is converted into grayscale. Next, feature extraction with a simple Min-Max normalization is applied. Obtained parameters are evaluated by classifiers and for each detected face, authors get posterior probability as the output of the classifier. Different approaches for face authentication are compared with each other using False Acceptance Rate (FAR), False Rejection Rate (FRR), Equal Error Rate (EER), Receiver Operating Characteristic (ROC) and Detection Error Tradeoff (DET) curves. The results are verified with AR Face Database and elaborated in a feature extraction and classifier design point of view. Best results were achieved by HOG feature for SVM classifier. Detailed results are listed in the text below.

Due to the enormous usage of the internet for transmission of data over a network, security and authenticity become major risks. Major challenges encountered in biometric system are the misuse of enrolled biometric templates stored in database server. To describe these issues various algorithms are implemented to deliver better protection to biometric traits such as physical (Face, fingerprint, Ear etc.) and behavioural (Gesture, Voice, tying etc.) by means of matching and verification process. In this work, biometric security system with fuzzy extractor and convolutional neural networks using face attribute is proposed which provides different choices for supporting cryptographic processes to the confidential data. The proposed system not only offers security but also enhances the system execution by discrepancy conservation of binary templates. Here Face Attribute Convolutional Neural Network (FACNN) is used to generate binary codes from nodal points which act as a key to encrypt and decrypt the entire data for further processing. Implementing Artificial Intelligence (AI) into the proposed system, automatically upgrades and replaces the previously stored biometric template after certain time period to reduce the risk of ageing difference while processing. Binary codes generated from face templates are used not only for cryptographic approach is also used for biometric process of enrolment and verification. Three main face data sets are taken into the evaluation to attain system performance by improving the efficiency of matching performance to verify authenticity. This system enhances the system performance by 8% matching and verification and minimizes the False Acceptance Rate (FAR), False Rejection Rate (FRR) and Equal Error Rate (EER) by 6 times and increases the data privacy through the biometric cryptosystem by 98.2% while compared to other work.

Today, numerous biometric systems have been suggested and created. Most used among them are face, fingerprint, and voice recognition. Each of them has their own advantages and disadvantages. Discussing about the automatic speaker verification (ASV), speech signal is a natural signal which one can easily obtain even from a telephone call. Research in this area is carried out from decades. This speaker verification technology has advanced in recent years and has become a genuine method for biometric systems. Using an adaptable Gaussian Mixture Model (GMM-UBM), a text-independent speaker verification (TISV) technique has been developed and compared with state-of-art I-vector based speaker recognition system in this research. Parameters for universal background model (UBM) are trained using the EM (Expectation maximization) and MAP adaptation method is used for training speaker models. Multiple false acceptance and false rejection rates are calculated by changing the threshold values for comparison. The results are shown in equal error rate (EER) for both the GMM-UBM, and I-vector based ASV systems, and the lowest EER is found to be 5.31% and 4.74%, respectively, after adjusting the threshold settings and number of Gaussian mixtures utilized.

This paper presents a robust user authentication system by gleaning raw mouse movement data. The data was collected using a publicly available tool called Recording User Input (RUI) from 23 subjects analyzed for three types of mouse actions - Mouse Move, Point-and-Click on Left or Right mouse button, and Drag-and-Drop. Samples are broken down to unit blocks comprising a certain number of actions and from each block seventy-four features are extracted to construct feature vectors. The proposed system was rigorously tested against public benchmark data. Experiment results generated by using the Support Vector Machine (SVM) classifier shows a False Rejection Rate (FRR) of 1.1594 % and a False Acceptance Rate (FAR) of 1.9053 % when the block size was set for 600 actions. After reducing dimensions using Principle Component Analysis (PCA), SVM classifier shows FRR of 1.2081 % and FAR of 2.3604 %. Compared with the existing methods based on mouse movements, our method shows significantly lower error rates, which we opine are viable enough to become an alternate to conventional authentication systems.

This paper presents a comparison between several features normalization methods, and a comparison between different types of Gaussian Mixture Model (GMM) based supervectors normalizations for robust Speaker Verification. We implemented the methods of normalizations as a part of speaker verification system using Support Vector Machine (SVM) classifier and GMM-based supervectors. When implementing the speaker recognition system, we used Mel Frequency Cepstral Coefficients (MFCC) feature extraction. A valid question is which features normalization to use, if any. We examine the most common methods of feature normalizations, such as: Feature Warping mapping, and Cepstral Mean Subtraction (CMS) normalization with and without variance normalization. These methods were compared to features without normalization at all, and to a basic [-1, 1] normalization. In addition, we applied few types of normalizations to the GMM-mean supervectors, in order to improve the performance of the SVM classifier. All comparisons of the speaker verification system had been done in terms of DET curve, EER (Equal Error Rate) and Min. DCF. The best results we achieved were on combined supervector normalizations of Universal Background Model (UBM) Standard Deviation (STD) and [-1, 1] normalization. The type of the MFCC normalization has no big influence on the verification performance. The best results were: EER about 5.0% and MIN. DCF of 0.02.

Replay, which is to playback a pre-recorded speech sample, presents a genuine risk to automatic speaker verification technology. In this study, we evaluate the vulnerability of text-dependent speaker verification systems under the replay attack using a standard benchmarking database, and also propose an anti-spoofing technique to safeguard the speaker verification systems. The key idea of the spoofing detection technique is to decide whether the presented sample is matched to any previous stored speech samples based a similarity score. The experiments conducted on the RSR2015 database showed that the equal error rate (EER) and false acceptance rate (FAR) increased from both 2.92 % to 25.56 % and 78.36 % respectively as a result of the replay attack. It confirmed the vulnerability of speaker verification to replay attacks. On the other hand, our proposed spoofing countermeasure was able to reduce the FARs from 78.36 % and 73.14 % to 0.06 % and 0.0 % for male and female systems, respectively, in the face of replay spoofing. The experiments confirmed the effectiveness of the proposed anti-spoofing technique.

This chapter aims to enhance the security of textual passwords by adding a new layer that involves an individual's typing rhythm without requiring additional devices. It discusses authentication methods and develops a textual password authentication system, which takes advantage of the fact that every user has a distinct way of typing. Twenty participants, including novice and expert users, assessed the proposed system. The measuring metrics used were the False Acceptance Rate (FAR) and False Rejection Rate (FRR). The flight time and typing rhythm were the biometric identification template. The results indicated low false rejection and acceptance rates, with the initial session registering 0.0 FRR and 0.1 FAR. Two weeks later, the second session recorded 0.25 FRR and 0.0 FAR. Using this model, users do not need to select a complex password they might forget. Instead, they can utilize a good rhythm different from their natural typing rhythm, making it challenging to guess.

We examined the false acceptance rate (FAR) and false rejection rate (FRR) of varying precision verification experimental designs. Analysis of variance was applied to derive the subcomponents of imprecision (ie, repeatability, between-run, between-day imprecision) for complex matrix experimental designs (day × run × replicate; day × run). For simple nonmatrix designs (1 day × multiple replicates or multiday × 1 replicate), ordinary standard deviations were calculated. The FAR and FRR in these different scenarios were estimated. The FRR increased as more samples were included in the precision experiment. The application of an upper verification limit, which seeks to cap FRR at 5% for multiple experiments, significantly increased the FAR. The FRR decreases as the observed imprecision increases relative to the claimed imprecision and when a greater number of days, runs, or replicates are included in the verification design. Increasing the number or days, runs, or replicates also reduces the FAR for between-day imprecision and repeatability. Design of verification experiments should incorporate the local availability of resources and analytical expertise. The largest imprecision component should be targeted with a greater number of measurements. Consideration of both FAR and FRR should be given when committing a platform into service.

Speaker recognition is an active research area that contains notable usage in biometric security and authentication system. Currently, there exist many well-performing models in the speaker recognition domain. However, most of the advanced models implement deep learning that requires GPU support for real-time speech recognition, and it is not suitable for low-end devices. In this paper, we propose a lightweight text-independent speaker recognition model based on random forest classifier. It also introduces new features that are used for both speaker verification and identification tasks. The proposed model uses human speech based timbral properties as features that are classified using random forest. Timbre refers to the very basic properties of sound that allow listeners to discriminate among them. The prototype uses seven most actively searched timbre properties, boominess, brightness, depth, hardness, roughness, sharpness, and warmth as features of our speaker recognition model. The experiment is carried out on speaker verification and speaker identification tasks and shows the achievements and drawbacks of the proposed model. In the speaker identification phase, it achieves a maximum accuracy of 78%. On the contrary, in the speaker verification phase, the model maintains an accuracy of 80% having an equal error rate (ERR) of 0.24.

Among all biometric technologies, iris recognition is the most accurate and high confidence authentication system. Due to the limitations in PCA-based system, modified principal component analysis (PCA)-based feature extraction is proposed in iris recognition system. In the proposed system, features are extracted using Gaussian kernel PCA (GKPCA) and classified using support vector machine (SVM). GKPCA and SVM algorithms are evaluated using CASIA V3 iris database. The performances are compared with the existing PCA-based system. The proposed system achieves 96.67% of accuracy for 256 features using GKPCA linear SVM. False acceptance rate (FAR) and false rejection rate (FRR) are 0 and 3 respectively for linear SVM. The results show that the proposed system performs accurate localisation of patterns even in non-ideal conditions.

Genuine person recognition systems are necessary for a system to identify or authorize the identification of a person requesting their services. Ensuring that only authorized users access the provided services is the aim of this proposed system. These systems are susceptible to impersonation attacks in the absence of strong personal recognition methods. Due to its traditional authentication technique (PIN), the ATM has suffered greatly over the years from PIN theft and other related ATM thefts. Therefore, in this paper, we proposed a multilayer (Typing and Touch) based authentication security arrangement to enhance the security and safety for the ATM and its users. The proposed system demonstrates a threelayered design structure. The first layer is the card verification module, which concentrates on the enrolment phase for ATM card. The second layer is the feature extraction and comparing of the typing pattern from the database server. The last layer compares the touch behavior continue till session end with the stored touch data in the database server. During the implementation phase, MySQL server 2010 served as the backend and Java was used for front-end development on Microsoft Windows 10. Based on the False Rejection Rate (FAR), False Acceptance Rate (FAR), and Equal Error Rate (EER) measurements, the application was evaluated. These metrics demonstrate the security and dependability of the suggested system for the authentication and verification of ATM customers.

This research paper presents a robust method for speaker verification in noisy environments. The noise is assumed to contaminate certain parts of the voice’s frequency spectrum. Therefore, the verification method is based on splitting the noisy speech into subsidiary bands then using a threshold to sense the existence of noise in a specific part of the spectrum, hence activating an adaptive filter in that part to track changes in noise’s characteristics and remove it. The decomposition is achieved using low complexity quadrature mirror filters QMF in three levels thus achieving four bands in a non-uniform that resembles human hearing perceptual. Speaker recognition is based on vector quantization VQ or template matching technique. Features are extracted from speaker’s voice using the normalized power in a similar way to the Mel-frequency cepstral coefficients. The performance of the proposed system is evaluated using 60 speakers subjected to five levels of signal to noise ratio SNR using total success rate TSR, false acceptance rate FAR, false rejection rate FRR and equal error rate. The proposed method showed higher recognition accuracy than existing methods in severe noise conditions.