Robust speech recognizer using multiclass SVM

Gavăt, Inge; Costache, Gabriel; Iancu, Claudia

doi:10.1109/neurel.2004.1416536

Cited by 3 publications

(3 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For our experiments on speech recognition, we use protocol 2, explained next. It is also used by other studies, such as that of Gavat et al [15].…”

Section: Xm2vts Databasementioning

confidence: 99%

“…The resulting vectors are called observation sequences and are then modeled with a Gaussian Mixture Model (GMM) and a Support Vector Machine (SVM) classifier for speech and speaker recognition, respectively. Recent studies report good performance with SVMs used as classifiers in recognition [36], [15], [8]. An SVM can provide a powerful discriminative classifier for finding models of the boundary between a speaker and impostors compared to traditional methods for speaker recognition such as GMMs [29] and artificial neural networks [14].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Synergy of Lip-Motion and Acoustic Features in Biometric Speech and Speaker Recognition

Faraj

Bigün

2007

IEEE Trans. Comput.

View full text Add to dashboard Cite

Abstract-This paper presents the scheme and evaluation of a robust audio-visual digit-and-speaker-recognition system using lip motion and speech biometrics. Moreover, a liveness verification barrier based on a person's lip movement is added to the system to guard against advanced spoofing attempts such as replayed videos. The acoustic and visual features are integrated at the feature level and evaluated first by a Support Vector Machine for digit and speaker identification and, then, by a Gaussian Mixture Model for speaker verification. Based on % 300 different personal identities, this paper represents, to our knowledge, the first extensive study investigating the added value of lip motion features for speaker and speech-recognition applications. Digit recognition and person-identification and verification experiments are conducted on the publicly available XM2VTS database showing favorable results (speaker verification is 98 percent, speaker identification is 100 percent, and digit identification is 83 percent to 100 percent).

show abstract

“…For our experiments on speech recognition, we use protocol 2, explained next. It is also used by other studies, such as that of Gavat et al [15].…”

Section: Xm2vts Databasementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Synergy of Lip-Motion and Acoustic Features in Biometric Speech and Speaker Recognition

Faraj

Bigün

2007

IEEE Trans. Comput.

View full text Add to dashboard Cite

show abstract

“…We propose two protocol setups for the XM2VTS database; protocol 1 is the well known Lausanne protocol [26], used for speaker identification and protocol 2 which is used for digit recognition. Protocol 2 is also suggested by other studies [14]. Protocol 1 -the training contains 225 subjects with 200 subjects as clients using and 25 subjects as impostors using sessions 1, 2 and 3.…”

Section: Xm2vts Databasementioning

confidence: 99%

Speaker and Digit Recognition by Audio-Visual Lip Biometrics

Faraj

Bigün

2007

Advances in Biometrics

View full text Add to dashboard Cite

Abstract. This paper proposes a new robust bi-modal audio visual digit and speaker recognition system by lip-motion and speech biometrics. To increase the robustness of digit and speaker recognition, we have proposed a method using speaker lip motion information extracted from video sequences with low resolution (128 ×128 pixels). In this paper we investigate a biometric system for digit recognition and speaker identification based using line-motion estimation with speech information and Support Vector Machines. The acoustic and visual features are fused at the feature level showing favourable results with digit recognition being 83% to 100% and speaker recognition 100% on the XM2VTS database.

show abstract

HIT-AVDB-II: A New Multi-view and Extreme Feature Cases Contained Audio-Visual Database for Biometrics

Lin¹,

Yao²,

Hong³

et al. 2008

Proceedings of the 11th Joint Conference on Information Sciences (JCIS)

View full text Add to dashboard Cite

For research on the proper law of audiovisual speech and biometrics technology, and evaluation of algorithms and systems, we construct a multi-language and multiview database HIT-AVDB-II with a corpus of various common and special sentences include Chinese and English poems, tongue twister, digits, Greek alphabet and music. The HIT-AVDB-II is ready to facilitate the investigation of multi-view biometrics technology and visual speech reading. HIT-AVDB-II contains formal and extreme feature cases for study. For fair comparison, we also establish related experiment protocols on view and intersection respectively. Further, we supplied one baseline speakeridentification recognition algorithm based on DCT to be compared.

show abstract

Robust speech recognizer using multiclass SVM

Cited by 3 publications

References 5 publications

Synergy of Lip-Motion and Acoustic Features in Biometric Speech and Speaker Recognition

Synergy of Lip-Motion and Acoustic Features in Biometric Speech and Speaker Recognition

Speaker and Digit Recognition by Audio-Visual Lip Biometrics

HIT-AVDB-II: A New Multi-view and Extreme Feature Cases Contained Audio-Visual Database for Biometrics

Contact Info

Product

Resources

About