| Paper: | SLP-P5.9 | ||
| Session: | Feature-based Robust Speech Recognition | ||
| Time: | Tuesday, May 16, 16:30 - 18:30 | ||
| Presentation: | Poster | ||
| Topic: | Speech and Spoken Language Processing: Feature-based Robust Speech Recognition (e.g., noise, etc) | ||
| Title: | PARAMETRIC NONLINEAR FEATURE EQUALIZATION FOR ROBUST SPEECH RECOGNITION | ||
| Authors: | Luz García, José C. Segura, Javier Ramírez, Angel de la Torre, Carmen Benítez, Universidad de Granada, Spain | ||
| Abstract: | A new front-end normalization algorithm using a parametric nonlinear transformation is proposed in this paper. It improves histogram equalization based transformations by finding a parametric expression of the nonlinear transformation. The new approach relies on a two Gaussian model for the probability distribution of the features, and on a Gaussian classifier to label the input frames as belonging to the speech or non-speech classes. The result is a more robust equalization, less dependent on the percentage of speech and non-speech frames. Recognition experiments on the AURORA 4 database have been performed and the effectiveness of the algorithm is analyzed in comparison with other linear and nonlinear feature equalization techniques. | ||
©2018 Conference Management Services, Inc. -||- email: webmaster@icassp2006.org -||- Last updated Friday, August 17, 2012