Speaker identification using multi-modal i-vector approach for varying length speech in voice interactive systems

Tiwari, Varun; Hashmi, Mohammad Farukh; Keskar, Avinash G.; Shivaprakash, N. C.

doi:10.1016/j.cogsys.2018.09.028

Cited by 12 publications

(6 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These parameters are usually adjusted repeatedly to decrease the cost function as the stochastic descent gradient decreases. Thus, for a given set of exercises, where x (i) is a given feature vector and y (i) corresponds to class (true), each hidden layer uses a nonlinear transformation function g to the result of the previous level [10]. This transformation takes into account the parameters W and b, which link one layer to the previous one, and provides the neuron activation values with the following ( 10), ( 11):…”

Section: Classifier System Deep Neural Networkmentioning

confidence: 99%

“…In the problem of text-dependent speaker recognition, a priori knowledge of the text content is used, while the hidden Markov models are more accurate than the models of Gaussian mixtures. In [10], some advantages of using a combination of speech recognition on the syllabic hidden Markov model and speaker-oriented recognition based on the Gaussian mixture model are shown. The systems under consideration, based on traditional Gaussian mixture models (GMM), have achieved satisfactory results for speaker recognition only when the speech length is large enough.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Development of security systems using DNN and i & x-vector classifiers

Mamyrbayev

Kydyrbekova

Alimhan

et al. 2021

EEJET

View full text Add to dashboard Cite

The widespread use of biometric systems entails increased interest from cybercriminals aimed at developing attacks to crack them. Thus, the development of biometric identification systems must be carried out taking into account protection against these attacks. The development of new methods and algorithms for identification based on the presentation of randomly generated key features from the biometric base of user standards will help to minimize the disadvantages of the above methods of biometric identification of users. We present an implementation of a security system based on voice identification as an access control key and a verification algorithm developed using MATLAB function blocks that can authenticate a person's identity by his or her voice. Our research has shown an accuracy of 90 % for this user identification system for individual voice characteristics. It has been experimentally proven that traditional MFCCs using DNN and i and x-vector classifiers can achieve good results. The paper considers and analyzes the most well-known approaches from the literature to the problem of user identification by voice: dynamic programming methods, vector quantization, mixtures of Gaussian processes, hidden Markov model. The developed software package for biometric identification of users by voice and the method of forming the user's voice standards implemented in the complex allows reducing the number of errors in identifying users of information systems by voice by an average of 1.5 times. Our proposed system better defines voice recognition in terms of accuracy, security and complexity. The application of the results obtained will improve the security of the identification process in information systems from various attacks.

show abstract

Section: Classifier System Deep Neural Networkmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Development of security systems using DNN and i & x-vector classifiers

Mamyrbayev

Kydyrbekova

Alimhan

et al. 2021

EEJET

View full text Add to dashboard Cite

show abstract

“…(2) The elderly or users with speech defects may encounter a series of problems in the home environment. The special speech recognition system developed can help them solve speech disorders and ensure their safety and health in life [33][34][35][36][37]. (3) In order to satisfy the users in the long distance voice interaction under special environment requirements, the designers of the analysis of the intelligent building system in the process of long distance voice interaction in which may exist problems, put forward the speech recognition system for long distances, and broke the voice interaction possible distance limit, to improve the user experience in the intelligence environment [38][39][40][41][42][43][44].…”

Section: The Voice Interactionmentioning

confidence: 99%

A Review of Smart Design Based on Interactive Experience in Building Systems

Zhang

et al. 2020

Sustainability

View full text Add to dashboard Cite

Smart building is the result of the penetration of information technology and control technology into traditional buildings, and is the future development direction of buildings. User-centric building smart design can achieve a sustainable life, and smart technology integration based on smart design can support sustainable development and improve user satisfaction, happiness, and overall quality of life. In intelligent design, researchers focus on the “people-oriented” approach, designed to bring users the ultimate interactive experience. Based on the interactive experience principle of smart design in the smart building system, this article classifies and summarizes intelligent design from the “five senses” interaction, including visual interaction, voice interaction, tactile interaction, cognitive interaction, and emotional interaction. We analyze the application of smart design in architecture and discuss how to embody the principles of user-centered interactive experience design in the process of smart design. This article provides a comprehensive and systematic literature review, clarifies the importance of the “people-oriented” approach in the smart design of buildings, and summarizes how to improve the sustainability of intelligent buildings from the perspective of a “people-oriented” approach. This paper proposes future research suggestions and directions and promotes the sustainable development of the smart building.

show abstract

“…Despite difficult acoustic settings, humans can extract speakerspecific speech qualities such as pitch, timbre, prosody and create latent speaker identity representations for recognition. With the aid of the state-of-the-art computational techniques, the notion of speaker identification has gained significant attention in the field of biometrics, access control systems that include voice dialing, interactive voice responses and others [2]. The efficiency of a typical speaker identification system solely depends on feature extraction and speaker modeling tasks [3].…”

Section: Introductionmentioning

confidence: 99%

“…The Gaussian mixture model-based approaches produce good results in speaker recognition task only when the speech lengths are sufficiently long. To tackle this problem, the authors in [2] proposed GMM based universal background model (GMM-UBM) to prepare multi-model i-vector speaker recognition system for short speech length.…”

Section: Introductionmentioning

confidence: 99%

A Novel Convolutional Neural Network Model for Automatic Speaker Identification From Speech Signals

Pandian,

Thirunavukarasu,

Kotei

2024

IEEE Access

View full text Add to dashboard Cite

Since voice communication is the most frequent form of human communication, it is necessary to identify each speaker in a discussion. The development of advanced computer processing units, deep learning techniques and publicly available datasets prompt the need for proposing novel automated speaker recognition systems. This study presents a novel Convolutional Neural Network (CNN) called Five Convolutional Blocks-CNN (5C-CNN) for automatic speaker identification from speech signals. The proposed 5C-CNN model comprises five Convolutional blocks and one dense block for identifying the speaker using their English speech audio. The proposed framework was trained with English audio speech signals collected from Ten individuals. The study also applied different audio augmentation techniques like white Gaussian noise, shifting, stretching, stretching with high frequency, altering speed, and changing pitch values to introduce diversity in the dataset. HyperBand tuning technique was used to optimize the training parameters of the proposed approach. The proposed 5C-CNN achieved a classification accuracy of 99.34% on the test data of the proposed person speech dataset. Also, the proposed approach performed well when tested against a benchmark dataset namely THUYG-20 for better generalization with an accuracy of 95.43%. Both experimental investigations clearly confirm the efficiency of the proposed technique in the audio classification task while comparing the existing approaches, and standard machine learning techniques.

show abstract

Speaker identification using multi-modal i-vector approach for varying length speech in voice interactive systems

Cited by 12 publications

References 23 publications

Development of security systems using DNN and i & x-vector classifiers

Development of security systems using DNN and i & x-vector classifiers

A Review of Smart Design Based on Interactive Experience in Building Systems

A Novel Convolutional Neural Network Model for Automatic Speaker Identification From Speech Signals

Contact Info

Product

Resources

About