Fig. 11From: A simulation study on optimal scores for speaker recognitionPerformance loss with data shift between the training and deployment. The x-axis represents the value shifted on all the dimensionsBack to article page