Discrimination parole/musique et étude de nouveaux paramètres et modèles pour un système d'identification du locuteur dans le contexte de conférences téléphoniques /

Ezzaidi, Hassan

doi:10.1522/17603685

Cited by 4 publications

(7 citation statements)

References 65 publications

(6 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Automatic Speaker Recognition (ASR) [21], [12] is a technology that identifies a person from their voice, after capturing a digital model of a person's voice, to determine the unique way of talk about a human being. The captured voice is transformed into a unique model which is stored in a database.…”

Section: Presentationmentioning

confidence: 99%

“…In recognition of the speaker [20], [21] we make the difference between identification and verification [23] of the speaker, depending on whether the problem is to verify that the analyzed voice corresponds to the person who is supposed to produce it, or that it is a question of determining who, among a predetermined number of speakers, produced the analyzed signal.…”

Section: Types Of Speaker Recognitionmentioning

confidence: 99%

“…Speaker recognition [21] or verification [23] consists of assigning an identity to a segment (or group of segments) of speech. This comparison is made using a system called automatic speaker recognition (ASR) [23], [24] whose function is to determine whether these voice signals have been produced by the same person.…”

Section: Principlementioning

confidence: 99%

“…To build a speaker verification system [21], [26] progress in various areas is used. So to arrive at an ideal and complete speaker verification system [30] This shows that the main objective to be achieved through this work is to carry out a speaker verification system, which allows us access to information in an efficient and optimal manner, based on the voice biometric feature "voice", as an authentication password.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Recombined and Hybridized Speaker Recognition System (SRS)

Abdelmajid¹

2020

IJATCSE

View full text Add to dashboard Cite

We started this report with a description of the organs constituting the speaker recognition system (SRS), so we immediately went on to explore the types of speaker recognition and the difficulties cited at this level. Next, we carried out a review of the main approaches to combinations and hybridizations developed over the past ten years in the field of the automatic speech recognition (ASR), by carrying out an experimental and comparative study of such an approach.Thus, we summarized the advantages and disadvantages presented by these methods of extraction and classification of acoustic characteristics, and their contributions to ASR.We have developed an approach to estimate a relevant speaker verification system, using a combined extractor, and a hybrid classifier, which allowed us to build our system of recognition.

show abstract

Section: Presentationmentioning

confidence: 99%

Section: Types Of Speaker Recognitionmentioning

confidence: 99%

Section: Principlementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Recombined and Hybridized Speaker Recognition System (SRS)

Abdelmajid¹

2020

IJATCSE

View full text Add to dashboard Cite

show abstract

“…The most powerful features currently used are the MFCCs (Mel Frequency Cepstral Coefficients) (Davis and Mermelstein 1980), the LPC (Linear Prediction Coding) (Slifka and Anderson 1995) and the PLP (Perceptual Linear Predictive) (Hermansky 1990). However, these features are very sensitive to speech signal variability under real-life conditions (Baudoin and Jardin 1993;Mary and Yegnanarayana 2008a;Ezzaidi 2002) and this causes a significant performance degradation of the ASR systems. The speech signal variability is mostly due to environmental factors (presence of noise) or to speaker characteristics (e.g.…”

mentioning

confidence: 99%

Robust Arabic speech recognition in noisy environments using prosodic features and formant

Amrous

Debyeche

Amrouche

2011

Int J Speech Technol

View full text Add to dashboard Cite

This paper investigates the contribution of formants and prosodic features such as pitch and energy in Arabic speech recognition under real-life conditions. Our speech recognition system based on Hidden Markov Models (HMMs) is implemented using the HTK Toolkit. The frontend of the system combines features based on conventional Mel-Frequency Cepstral Coefficient (MFFC), prosodic information and formants. The experiments are performed on the ARADIGIT corpus which is a database of Arabic spoken words. The obtained results show that the resulting multivariate feature vectors, in noisy environment, lead to a significant improvement, up to 27%, in word accuracy relative the word accuracy obtained from the state-of-the-art MFCCbased system.

show abstract