A comparison of approaches for modeling prosodic features in speaker recognition | IEEE Conference Publication | IEEE Xplore