EUROSPEECH 2001 Scandinavia
7th European Conference on Speech Communication and Technology

Aalborg, Denmark
September 3-7, 2001


Robust ASR front-end using spectral-based and discriminant features: experiments on the Aurora tasks

M. Carmen Benitez (1), Lukas Burget (2), Barry Chen (1), Stephane Dupont (1), Hari Garudadri (3), Hynek Hermansky (2), Pratibha Jain (2), Sachin Kajarekar (2), Nelson Morgan (1), Sunil Sivadas (2)

(1) International Computer Science Institute, USA
(2) Oregon Graduate Institute Of Science and Technology, USA
(3) Qualcomm Inc., USA

This paper describes an automatic speech recognition front-end that combines low-level robust ASR feature extraction techniques, and higher-level linear and non-linear feature transformations. The low-level algorithms use data-derived filters, mean and variance normalization of the feature vectors, and dropping of noise frames. The feature vectors are then linearly transformed using Principal Components Analysis (PCA). An Artificial Neural Network (ANN) is also used to compute features that are useful for classification of speech sounds. It is trained for phoneme probability estimation on a large corpus of noisy speech. These transformations lead to two feature streams whose vectors are concatenated and then used for speech recognition. This method was tested on the set of speech corpora used for the Aurora evaluation. Using the feature stream generated without the ANN yields an overall 41% reduction of the error rate over Mel-Frequency Cepstral Coefficients (MFCC) reference features. Adding the ANN stream further reduces the error rate yielding a 46% reduction over the reference features.

Full Paper

Bibliographic reference.  Benitez, M. Carmen / Burget, Lukas / Chen, Barry / Dupont, Stephane / Garudadri, Hari / Hermansky, Hynek / Jain, Pratibha / Kajarekar, Sachin / Morgan, Nelson / Sivadas, Sunil (2001): "Robust ASR front-end using spectral-based and discriminant features: experiments on the Aurora tasks", In EUROSPEECH-2001, 429-432.