Fig. 1From: Introducing phonetic information to speaker embedding for speaker verificationThe x-vector architecture. This architecture can be partitioned into frame- and segment-level sub-networksBack to article page