Fig. 9From: A simulation study on optimal scores for speaker recognitionPerformance loss with between-class variance mismatch between the training and deployment. The x-axis represents the distortion factor α on the between-class variancesBack to article page