Iterative Constrained MLLR Approach for Speaker Adaptation

In this paper an effective technique for speaker adaptation on the feature domain is presented. This technique starts from the well known maximum-likelihood linear regression (MLLR) auxiliary function to obtain the constrained MLLR transformation in an iterative fashion. The proposed approach is particularly suitable to be implemented on the client side of a distributed speech recognition scheme, due to the reduced number of iterations required to reach convergence. Extensive experimentation using the CMU Sphinx 4 ASR system along with a preliminarily trained speaker-independent acoustic model for the Italian language, in a setting designed for large-vocabulary continuous speech recognition, demonstrates the effectiveness of the approach even with small amounts of adaptation data.