Figure 8From: An evolutionary feature synthesis approach for content-based audio retrievalThe AP performances obtained by consecutive SLP classifiers. In general, the results cannot be further improved by concatenating the SLPs, except with the segment features for the first few classifiers.Back to article page