Figure 9From: Integrated exemplar-based template matching and statistical modeling for continuous speech recognitionPhone accuracies (percent) versus the percentage of template representatives for MDTS (LLR) and MLTS (KL). For MDTS and MLTS with their respectively compatible distances, when less templates were used, worse performance was obtained. MLTS is more robust to using a small percentage of template representatives, and the selection percentage of 20% is a reasonable compromise between accuracy performance and computation and storage cost.Back to article page