Mokbel, C., Mauuary, L., Karray, L., Jouvet, D., Monn, J., Simonin, J., and Bartkova, K., "Towards improving ASR robustness for PSN and GSM telephone applications ", Speech Comm., 23, 141 -- 159, 1997.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....Recognition, Histogram Equalization, Vector Taylor Series approach, Mel Frequency Cepstral Coe#cients, Cepstral Mean Normalization. I. Introduction The noise is one of the most important problems to be considered in Automatic Speech Recognition systems working in real conditions [1] 2] 3] [4] [5] 6] 7] Currently, robust speech recognition is one of the research lines focusing most of the attention in the context of speech technologies and their applications [8] 9] 10] 11] 12] 13] The noise has two main e#ects over the representation of the speech signal. On one hand, it ....

C. Mokbel, L. Mauuary, L. Karray, D. Jouvet, J. Monne, J. Simonin, and K. Bartkova. Towards improving ASR robustness for PSN and GSM telephone applications. Speech Communication, 23(1-2):141--160, October 1997.

....ML and then adapts it for each client separately using his own training material. This general methodology have already been applied in other domains than speaker veri cation, such as Automatic Speech Recognition (ASR) where it generally yielded very good performance (see for instance [7] or [14]) Some authors have even studied the case where the adaptation is done incrementally in parallel with the real use of the system [3] Many di erent strategies can be used to adapt a generic model using a small training set. In this chapter, we present two of the most used adaptation methods in ....

C. Mokbel, L. Mauuary, L. Karray, D. Jouvet, J. Monn, J. Simonin, and K. Bartkova. Towards improving ASR robustness for PSN and GSM telephone applications. Speech Communication, 23(1):141159, 1997.

.... performance, see e.g. the work performed by Euler and Zinke (1993) Lilly and Palival (1996) or Tucker et al. 1999) The investigations aim to develop recognition systems which are robust towards the specific impairment, e.g. by using preprocessing and adaptation techniques (Mokbel et al. 1993; Mokbel et al. 1997), or by training acoustic models with impaired speech data (e.g. Puel and Andr Obrecht, 1997) Robust HMM architectures have also been proposed, e.g. for impairments which are to be encountered in GSM cellular networks (interruptions and impulsive noise) by Karray et al. 1998) Making use of ....

Mokbel, C., Mauuary, L., Karray, L., Jouvet, D., Monn, J., Simonin, J. and Bartkova, K., 1997. Towards Improving ASR Robustness for PSN and GSM Telephone Applications. Speech Communication 23, 141-159.

....communication and its associated background additive noise. The signal is telephony band limited and noise compensation is performed locally prior to communication channel processing. Non linear spectral subtraction (NLSS) has been applied both speech enhancement [6, 7] and noise compensation [1, 2, 8, 9] and uses a spectral estimation of the noise obtained during speech pauses. This is then used to compensate a speech signal degraded by additive noise. Here the experimental study aims to quantify the e ectiveness of various NLSS techniques reported in the literature. Short term spectral estimates ....

C. Mokbel, L. Mauuary, L Karray, D. Jouvet, J. Monn, J. Simonin, and K. Bartkova. Towards improving ASR robustness for PSN and GSM telephone applications. Speech Communication, 23:141159, October 1997.

No context found.

Mokbel, C., Mauuary, L., Karray, L., Jouvet, D., Monn, J., Simonin, J., and Bartkova, K., "Towards improving ASR robustness for PSN and GSM telephone applications ", Speech Comm., 23, 141 -- 159, 1997.

No context found.

C. Mokbel, L. Mauuary, L. Karray, D. Jouvet, J. Monne, J. Simonin, and K. Bartkova, "Towards improving ASR robustness for PSN and GSM telephone applications," Speech communication, vol. 3, pp. 141--159, May 1997.

....have been successfully applied in several domains of speech and speaker recognition. These techniques allow supervised or unsupervised adaptation of a recognition system to a particular condition of use, e.g. a particular speaker or a speci c environment by adjusting speaker model parameters [3] 6][8]. In that case, the initial model is generally estimated using a large amount of data. Adaptation techniques are also used in the context of speaker veri cation, for estimating a speaker model as an adapted version of a speaker independent model [10] The work presented here has been carried out ....

C. Mokbel, L. Mauuary, L. Karray, D. Jouvet, J. Monne, J. Simonin, K. Bartkova, Towards Improving ASR Robustness for PSN and GSM Telephone Applications, Speech Communication, V. 23, n1, pp. 141-159, Oct. 1997.

No context found.

C. Mokbel, et al., "Towards improving ASR robustness for PSN and GSM telephone applications," Speech Communication, vol. 23, nos. 1-2, pp. 141-159, Oct. 1997.

No context found.

C. Mokbel, L. Mauuary, L. Karray, D. Jouvet, J Monne, J. Simonin, and K. Bartkova. Towards improving asr robustness for psn and gsm telephone applications. Speech Communication, 23:141-159, 1997.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC