Robust speaker recognition using MAP estimation of additive noise in i-vectors space

Abstract

In the last few years, the use of i-vectors along with a generative back-end has become the new standard in speaker recognition. An i-vector is a compact representation of a speaker utterance extracted from a low dimensional total variability subspace. Although current speaker recognition systems achieve very good results in clean training and test conditions, the performance degrades considerably in noisy environments. The compensation of the noise effect is actually a research subject of major importance. As far as we know, there was no serious attempt to treat the noise problem directly in the i-vectors space without relying on data distributions computed on a prior domain. This paper proposes a full-covariance Gaussian modeling of the clean i-vectors and noise distributions in the i-vectors space then introduces a technique to estimate a clean i-vector given the noisy version and the noise density function using MAP approach. Based on NIST data, we show that it is possible to improve up to 60% the baseline system performances. A noise adding tool is used to help simulate a real-world noisy environment at different signal-to-noise ratio levels.