In this paper, we propose a hybrid system based on a modified statistical GMM voice conversion algorithm for improving the recognition of esophageal speech. This hybrid system aims to compensate for the distorted information present in the esophageal acoustic features by using a voice conversion method. The esophageal speech is converted into a “target” laryngeal speech using an iterative statistical estimation of a transformation function. We did not apply a speech synthesizer for reconstructing the converted speech signal, given that the converted Mel cepstral vectors are used directly as input of our speech recognition system. Furthermore the feature vectors are linearly transformed by the HLDA (heteroscedastic linear discriminant analysis) method to reduce their size in a smaller space having good discriminative properties. The experimental results demonstrate that our proposed system provides an improvement of the phone recognition accuracy with an absolute increase of 3.40 % when compared with the phone recognition accuracy obtained with neither HLDA nor voice conversion.

No abstract

In this paper, we propose a simple and fast method for evaluating the pathological voice (esophageal) by applying the continuous speech recognition in a speaker dependent mode, on our own database of the pathological voice, we call FPSD (French Pathological Speech Database). The recognition system used is implemented using the HTK platform, based on HMM/GMM monophone models. The acoustic vectors are linearly transformed by the HLDA (Heteroscedastic Linear Discriminant Analysis) method to reduce their size in a smaller space with good discriminative properties. The obtained phone recognition rate (63.59 %) is very promising when we know that esophageal voice contains unnatural sounds, difficult to understand.