Rhythm Features for Speaker Identification

Nick Mehlman, Thomas Thebaud, Dani Byrd +1

While deep learning models have demonstrated robust performance in speaker recognition tasks, they primarily rely on low-level audio features learned empirically from spectrograms…