Speaker Verification from Coded Telephone Speech Using Stochastic Feature Transformation and Handset Identification

Yu, E.W.M.; Mak, Man‐Wai; Kung, Sun‐Yuan

doi:10.1007/3-540-36228-2_74

Cited by 9 publications

(3 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We used a GSM speech codec to transcode the HTIMIT corpus [9] and applied the resulting transcoded speech in a speaker verification experiment similar to [10] and [11]. HTIMIT was obtained by playing a subset of the TIMIT corpus through 9 different telephone handsets and one Sennheizer head-mounted microphone.…”

Section: Speaker Verification Experimentsmentioning

confidence: 99%

Probabilistic Fusion of Sorted Score Sequences for Robust Speaker Verification

Cheung

Mak

Kung

2005

Intelligent Multimedia Processing With Soft Computing

Self Cite

View full text Add to dashboard Cite

Abstract. Fusion techniques have been widely used in multi-modal biometric authentication systems. While these techniques are mainly applied to combine the outputs of modality-dependent classifiers, they can also be applied to fuse the decisions or scores from a single modality. The idea is to consider the multiple samples extracted from a single modality as independent but coming from the same source. In this chapter, we propose a single-source, multi-sample data-dependent fusion algorithm for speaker verification. The algorithm is data-dependent in that the fusion weights are dependent on the verification scores and the prior score statistics of claimed speakers and background speakers. To obtain the best out of the speaker's scores, scores from multiple utterances are sorted before they are probabilistically combined. Evaluations based on 150 speakers from a GSM-transcoded corpus are presented. Results show that data-dependent fusion of speaker's scores is significantly better than the conventional score averaging approach. It was also found that the proposed fusion algorithm can be further enhanced by sorting the score sequences before they are probabilistically combined.

show abstract

Section: Speaker Verification Experimentsmentioning

confidence: 99%

Probabilistic Fusion of Sorted Score Sequences for Robust Speaker Verification

Cheung

Mak

Kung

2005

Intelligent Multimedia Processing With Soft Computing

Self Cite

View full text Add to dashboard Cite

show abstract

“…The speech bit rate and mismatched training and testing conditions are the main obstacles for efficient speaker ID systems [2]. Observable strong correlation between speech bit rate and identification accuracy stimulates applying compensation methods based on features transformation [4], [5]. Another approach [3] exploits codebook and LPC compensation techniques to reduce distortion and demand modifications to the speech coder.…”

Section: Introductionmentioning

confidence: 99%

“…The impact of speech coding effects on speaker ID has been investigated already in few works [2]- [4]. The speech bit rate and mismatched training and testing conditions are the main obstacles for efficient speaker ID systems [2].…”

Section: Introductionmentioning

confidence: 99%

Speech coding influence on features dedicated to speaker identification

Maka¹,

Bonikowski²

2008

2008 International Conference on Signals and Electronic Systems

View full text Add to dashboard Cite

The analysis of three sets of feature vectors used in speaker identification (ID) systems for speech signals received in encoding-decoding process with AMR, SPEEX and MELP coders has been presented. We have analyzed feature sets for various speech coding bit rates using SVM-based speaker ID system. The results were compared with identification accuracy obtained with vectors where fundamental frequency was an additional feature. Performed experiments show that such feature contributes better identification accuracy for coded speech than uncoded one in most cases.

show abstract

Cluster-Dependent Feature Transformation for Telephone-Based Speaker Verification

Tsang

Mak

Kung

2003

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Abstract. This paper presents a cluster-based feature transformation technique for telephone-based speaker verification when labels of the handset types are not available during the training phase. The technique combines a cluster selector with cluster-dependent feature transformations to reduce the acoustic mismatches among different handsets. Specifically, a GMM-based cluster selector is trained to identify the cluster that best represents the handset used by a claimant. Handset distorted features are then transformed by cluster-specific feature transformation to remove the acoustic distortion before being presented to the clean speaker models. Experimental results show that cluster-dependent feature transformation with number of clusters larger than the actual number of handsets can achieve a performance level very close to that achievable by the handset-based transformation approaches.

show abstract

Speaker Verification from Coded Telephone Speech Using Stochastic Feature Transformation and Handset Identification

Cited by 9 publications

References 12 publications

Probabilistic Fusion of Sorted Score Sequences for Robust Speaker Verification

Probabilistic Fusion of Sorted Score Sequences for Robust Speaker Verification

Speech coding influence on features dedicated to speaker identification

Cluster-Dependent Feature Transformation for Telephone-Based Speaker Verification

Contact Info

Product

Resources

About