Figure 6From: Integrated exemplar-based template matching and statistical modeling for continuous speech recognitionPhone accuracies based on different methods. Comparison on phone accuracies (percent) from the HMM baseline and the template-matching-based lattice rescoring with the local distances of Mahalanobis, NLL, LLR, and KL, where in the last three cases 1GMM was used in labeling each frame vector.Back to article page