scholarly journals Robust Deep Speaker Recognition: Learning Latent Representation with Joint Angular Margin Loss

2020 ◽  
Vol 10 (21) ◽  
pp. 7522
Author(s):  
Labib Chowdhury ◽  
Hasib Zunair ◽  
Nabeel Mohammed
Keyword(s):  
Speaker Recognition ◽  
Speaker Identification ◽  
Biometric Authentication ◽  
Convolutional Network ◽  
Performance Improvements ◽  
Design Choice ◽  
Sinc Functions ◽  
Recognition Learning ◽  
Prior Models ◽  
Potential Successor

Speaker identification is gaining popularity, with notable applications in security, automation, and authentication. For speaker identification, deep-convolutional-network-based approaches, such as SincNet, are used as an alternative to i-vectors. Convolution performed by parameterized sinc functions in SincNet demonstrated superior results in this area. This system optimizes softmax loss, which is integrated in the classification layer that is responsible for making predictions. Since the nature of this loss is only to increase interclass distance, it is not always an optimal design choice for biometric-authentication tasks such as face and speaker recognition. To overcome the aforementioned issues, this study proposes a family of models that improve upon the state-of-the-art SincNet model. Proposed models AF-SincNet, Ensemble-SincNet, and ALL-SincNet serve as a potential successor to the successful SincNet model. The proposed models are compared on a number of speaker-recognition datasets, such as TIMIT and LibriSpeech, with their own unique challenges. Performance improvements are demonstrated compared to competitive baselines. In interdataset evaluation, the best reported model not only consistently outperformed the baselines and current prior models, but also generalized well on unseen and diverse tasks such as Bengali speaker recognition.

scholarly journals Speaker Recognition using Random Forest

2021 ◽  
Vol 37 ◽  
pp. 01022
Author(s):  
K Khadar Nawas ◽  
Manish Kumar Barik ◽  
A Nayeemulla Khan
Keyword(s):  
Random Forest ◽  
Speaker Recognition ◽  
Speaker Identification ◽  
Promising Result ◽  
Identification System ◽  
Biometric Authentication ◽  
Human Machine Interaction ◽  
Accuracy Level ◽  
Authentication Security ◽  
Machine Interaction

Speaker identification has become a mainstream technology in the field of machine learning that involves determining the identity of a speaker from his/her speech sample. A person’s speech note contains many features that can be used to discriminate his/her identity. A model that can identify a speaker has wide applications such as biometric authentication, security, forensics and human-machine interaction. This paper implements a speaker identification system based on Random Forest as a classifier to identify the various speakers using MFCC and RPS as feature extraction techniques. The output obtained from the Random Forest classifier shows promising result. It is observed that the accuracy level is significantly higher in MFCC as compared to the RPS technique on the data taken from the well-known TIMIT corpus dataset.

New Feature Vectors using GFCC for Speaker Identification

2018 ◽  
Vol 6 (8) ◽  
pp. 243
Author(s):  
A. Nagesh
Keyword(s):  
Speaker Recognition ◽  
Speaker Identification ◽  
Signal To Noise Ratio ◽  
Main Idea ◽  
Extraction Methods ◽  
Identification System ◽  
Identification Performance ◽  
Feature Vectors ◽  
Overall Performance ◽  
New Feature

The feature vectors of speaker identification system plays a crucial role in the overall performance of the system. There are many new feature vectors extraction methods based on MFCC, but ultimately we want to maximize the performance of SID system.  The objective of this paper to derive Gammatone Frequency Cepstral Coefficients (GFCC) based a new set of feature vectors using Gaussian Mixer model (GMM) for speaker identification. The MFCC are the default feature vectors for speaker recognition, but they are not very robust at the presence of additive noise. The GFCC features in recent studies have shown very good robustness against noise and acoustic change. The main idea is  GFCC features based on GMM feature extraction is to improve the overall speaker identification performance in low signal to noise ratio (SNR) conditions.

scholarly journals Performance Evaluation of Mel and Bark Scale based Features for Text-Independent Speaker Identification

2019 ◽  
Vol 8 (11) ◽  
pp. 3734-3738
Keyword(s):  
Speaker Recognition ◽  
Filter Bank ◽  
Work Performance ◽  
Speaker Identification ◽  
Recognition Rate ◽  
Identification System ◽  
Bank Structure ◽  
Speech Features ◽  
Recognition Systems ◽  
Human Ear

The performance of Mel scale and Bark scale is evaluated for text-independent speaker identification system. Mel scale and Bark scale are designed according to human auditory system. The filter bank structure is defined using Mel and Bark scales for speech and speaker recognition systems to extract speaker specific speech features. In this work, performance of Mel scale and Bark scale is evaluated for text-independent speaker identification system. It is found that Bark scale centre frequencies are more effective than Mel scale centre frequencies in case of Indian dialect speaker databases. Mel scale is defined as per interpretation of pitch by human ear and Bark scale is based on critical band selectivity at which loudness becomes significantly different. The recognition rate achieved using Bark scale filter bank is 96% for AISSMSIOIT database and 95% for Marathi database.

scholarly journals Comparison of feature extraction and normalization methods for speaker recognition using grid-audiovisual database

2020 ◽  
Vol 18 (2) ◽  
pp. 782
Author(s):  
Musab T. S. Al-Kaltakchi ◽  
Haithem Abd Al-Raheem Taha ◽  
Mohanad Abd Shehab ◽  
Mohamed A.M. Abdullah
Keyword(s):  
Feature Extraction ◽  
Speaker Recognition ◽  
Speaker Identification ◽  
Gaussian Mixture ◽  
Identification Accuracy ◽  
Identification System ◽  
Good Representation ◽  
Mel Frequency Cepstral Coefficients ◽  
Normalization Methods ◽  
Cepstral Coefficients

<p><span lang="EN-GB">In this paper, different feature extraction and feature normalization methods are investigated for speaker recognition. With a view to give a good representation of acoustic speech signals, Power Normalized Cepstral Coefficients (PNCCs) and Mel Frequency Cepstral Coefficients (MFCCs) are employed for feature extraction. Then, to mitigate the effect of linear channel, Cepstral Mean-Variance Normalization (CMVN) and feature warping are utilized. The current paper investigates Text-independent speaker identification system by using 16 coefficients from both the MFCCs and PNCCs features. Eight different speakers are selected from the GRID-Audiovisual database with two females and six males. The speakers are modeled using the coupling between the Universal Background Model and Gaussian Mixture Models (GMM-UBM) in order to get a fast scoring technique and better performance. The system shows 100% in terms of speaker identification accuracy. The results illustrated that PNCCs features have better performance compared to the MFCCs features to identify females compared to male speakers. Furthermore, feature wrapping reported better performance compared to the CMVN method. </span></p>

scholarly journals The processing of intimately familiar and unfamiliar voices: Specific neural responses of speaker recognition and identification

PLoS ONE ◽  
2021 ◽  
Vol 16 (4) ◽  
pp. e0250214
Author(s):  
Julien Plante-Hébert ◽  
Victor J. Boucher ◽  
Boutheina Jemel
Keyword(s):  
Speaker Recognition ◽  
Time Window ◽  
Speaker Identification ◽  
Early Recognition ◽  
Voice Recognition ◽  
Event Related Potentials ◽  
Neural Responses ◽  
Speech Stimuli ◽  
Related Potentials ◽  
Post Onset

Research has repeatedly shown that familiar and unfamiliar voices elicit different neural responses. But it has also been suggested that different neural correlates associate with the feeling of having heard a voice and knowing who the voice represents. The terminology used to designate these varying responses remains vague, creating a degree of confusion in the literature. Additionally, terms serving to designate tasks of voice discrimination, voice recognition, and speaker identification are often inconsistent creating further ambiguities. The present study used event-related potentials (ERPs) to clarify the difference between responses to 1) unknown voices, 2) trained-to-familiar voices as speech stimuli are repeatedly presented, and 3) intimately familiar voices. In an experiment, 13 participants listened to repeated utterances recorded from 12 speakers. Only one of the 12 voices was intimately familiar to a participant, whereas the remaining 11 voices were unfamiliar. The frequency of presentation of these 11 unfamiliar voices varied with only one being frequently presented (the trained-to-familiar voice). ERP analyses revealed different responses for intimately familiar and unfamiliar voices in two distinct time windows (P2 between 200–250 ms and a late positive component, LPC, between 450–850 ms post-onset) with late responses occurring only for intimately familiar voices. The LPC present sustained shifts, and short-time ERP components appear to reflect an early recognition stage. The trained voice equally elicited distinct responses, compared to rarely heard voices, but these occurred in a third time window (N250 between 300–350 ms post-onset). Overall, the timing of responses suggests that the processing of intimately familiar voices operates in two distinct steps of voice recognition, marked by a P2 on right centro-frontal sites, and speaker identification marked by an LPC component. The recognition of frequently heard voices entails an independent recognition process marked by a differential N250. Based on the present results and previous observations, it is proposed that there is a need to distinguish between processes of voice “recognition” and “identification”. The present study also specifies test conditions serving to reveal this distinction in neural responses, one of which bears on the length of speech stimuli given the late responses associated with voice identification.

scholarly journals SPEAKER IDENTIFICATION MENGGUNAKAN TRANSFORMASI WAVELET DISKRIT DAN JARINGAN SARAF TIRUAN BACK-PROPAGATION

2008 ◽  
Vol 2 (1) ◽  
pp. 1
Author(s):  
Anny Tandyo ◽  
Martono Martono ◽  
Adi Widyatmoko
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Speaker Recognition ◽  
Speaker Identification ◽  
Back Propagation ◽  
Identification Accuracy ◽  
Identification System ◽  
Accuracy Rate ◽  
Discrete Transformation ◽  
Artificial Neural

Article discussed a speaker identification system. Which was a part of speaker recognition. The system identified asubject based on the voice from a group of pattern had been saved before. This system used a wavelet discrete transformationas a feature extraction method and an artificial neural network of back-propagation as a classification method. The voiceinput was processed by the wavelet discrete transformation in order to obtain signal coefficient of low frequency as adecomposition result which kept voice characteristic of everyone. The coefficient then was classified artificial neural networkof back-propagation. A system trial was conducted by collecting voice samples directly by using 225 microphones in nonsoundproof rooms; contained of 15 subjects (persons) and each of them had 15 voice samples. The 10 samples were used as atraining voice and 5 others as a testing voice. Identification accuracy rate reached 84 percent. The testing was also done onthe subjects who pronounced same words. It can be concluded that, the similar selection of words by different subjects has noinfluence on the accuracy rate produced by system.Keywords: speaker identification, wavelet discrete transformation, artificial neural network, back-propagation.

Speaker Verification and Identification

2010 ◽  
pp. 264-289
Author(s):  
Minho Jin ◽  
Chang D. Yoo
Keyword(s):  
Feature Extraction ◽  
Speaker Recognition ◽  
Speaker Identification ◽  
Speaker Verification ◽  
Gaussian Mixture ◽  
Recognition System ◽  
Essential Elements ◽  
Verification System ◽  
Biometric Characteristic ◽  
Speaker Modeling

A speaker recognition system verifies or identifies a speaker’s identity based on his/her voice. It is considered as one of the most convenient biometric characteristic for human machine communication. This chapter introduces several speaker recognition systems and examines their performances under various conditions. Speaker recognition can be classified into either speaker verification or speaker identification. Speaker verification aims to verify whether an input speech corresponds to a claimed identity, and speaker identification aims to identify an input speech by selecting one model from a set of enrolled speaker models. Both the speaker verification and identification system consist of three essential elements: feature extraction, speaker modeling, and matching. The feature extraction pertains to extracting essential features from an input speech for speaker recognition. The speaker modeling pertains to probabilistically modeling the feature of the enrolled speakers. The matching pertains to matching the input feature to various speaker models. Speaker modeling techniques including Gaussian mixture model (GMM), hidden Markov model (HMM), and phone n-grams are presented, and in this chapter, their performances are compared under various tasks. Several verification and identification experimental results presented in this chapter indicate that speaker recognition performances are highly dependent on the acoustical environment. A comparative study between human listeners and an automatic speaker verification system is presented, and it indicates that an automatic speaker verification system can outperform human listeners. The applications of speaker recognition are summarized, and finally various obstacles that must be overcome are discussed.

Voice Biometrics

2018 ◽  
pp. 360-389
Author(s):  
Saleema A. ◽  
Sabu M. Thampi
Keyword(s):  
Present State ◽  
Speaker Identification ◽  
Fog Computing ◽  
Background Information ◽  
Biometric Authentication ◽  
Biometric Technology ◽  
Extensive Survey ◽  
Authentication System ◽  
State Of Affairs ◽  
Voice Biometrics

Biometric technology is spearheading the existing authentication methods in the IoT. Considering the balance between security and convenience, voice biometrics seems to be the most logical biometric technologies to be used. The authors present an extensive survey to identify, analyze, and compare various methods and algorithms for the different phases in the process of speaker identification/recognition, which is the part and parcel in voice biometrics. The chapter is intended to provide essential background information to those interested in learning or planning to design voice authentication systems. The chapter highlights the need for a biometric authentication system, the reason why we prefer voice, its present state of affairs, and its scope with fog computing to be used in IoT.

An Experimental Comparison of Modeling Techniques and Combination of Speaker – Specific Information from Different Languages for Multilingual Speaker Identification

2016 ◽  
Vol 25 (4) ◽  
pp. 529-538
Author(s):  
H.S. Jayanna ◽  
B.G. Nagaraja
Keyword(s):  
Speaker Recognition ◽  
English Language ◽  
Speaker Identification ◽  
Poor Performance ◽  
Gaussian Mixture ◽  
Experimental Comparison ◽  
Identification System ◽  
Specific Information ◽  
Self Organizing Map ◽  
Modeling Techniques

AbstractMost of the state-of-the-art speaker identification systems work on a monolingual (preferably English) scenario. Therefore, English-language autocratic countries can use the system efficiently for speaker recognition. However, there are many countries, including India, that are multilingual in nature. People in such countries have habituated to speak multiple languages. The existing speaker identification system may yield poor performance if a speaker’s train and test data are in different languages. Thus, developing a robust multilingual speaker identification system is an issue in many countries. In this work, an experimental evaluation of the modeling techniques, including self-organizing map (SOM), learning vector quantization (LVQ), and Gaussian mixture model-universal background model (GMM-UBM) classifiers for multilingual speaker identification, is presented. The monolingual and crosslingual speaker identification studies are conducted using 50 speakers of our own database. It is observed from the experimental results that the GMM-UBM classifier gives better identification performance than the SOM and LVQ classifiers. Furthermore, we propose a combination of speaker-specific information from different languages for crosslingual speaker identification, and it is observed that the combination feature gives better performance in all the crosslingual speaker identification experiments.

scholarly journals A CNN based Speaker Recognition System using an Alternate Bone Microphone

2019 ◽  
Vol 9 (2) ◽  
pp. 4224-4227
Keyword(s):  
Speaker Recognition ◽  
Input Data ◽  
Speaker Identification ◽  
Recognition System ◽  
The Other ◽  
Identification System ◽  
Video Recorder ◽  
State Of Art

State-of-art speaker recognition system uses acoustic microphone speech to identify/verify a speaker. The multimodal speaker recognition system includes modality of input data recorded using sources like acoustics mic,array mic ,throat mic, bone mic and video recorder. In this paper we implemented a multi-modal speaker identification system with three modality of speech as input, recorded from different microphones like air mic, throat mic and bone mic . we propose and claim an alternate way of recording the bone speech using a throat microphone and the results of a implemented speaker recognition using CNN and spectrogram is presented. The obtained results supports our claim to use the throat microphone as suitable mic to record the bone conducted speech and the accuracy of the speaker recognition system with signal speech recorded from air microphone get improved about 10% after including the other modality of speech like throat and bone speech along with the air conducted speech.

Sign in / Sign up

Export Citation Format

Share Document