Title: Histogram Based Normalization in the Acoustic Feature Space

Authors: Sirko Molau, Hermann Ney, Michael Pitz

Abstract: We describe a technique called histogram normalization that aims at normalizing feature space distributions at different stages in the signal analysis front-end, namely the log-compressed filterbank vectors, cepstrum 
coefficients, and LDA-transformed acoustic vectors. Best results are obtained at the filterbank, and in most cases there is a minor additional gain when 
normalization is applied sequentially at different stages.
We show that histogram normalization performs best if applied both in training and recognition, and that smoothing the target histogram obtained on the training data is also helpful.
On the VerbMobil II corpus, a German large-vocabulary conversational speech recognition task, we achieve an overall reduction in word error rate
of about 10% relative.

a01sm007.ps a01sm007.pdf