This paper proposes an algorithm to perform privacy-preserving (PP) speaker recognition using Gaussian mixture models (GMM). We consider a scenario where the users have to enrol their voice biometric with the third-party service providers to access …