2017-Monaural Speech Speakers Separation

  • To separate different speakers’ voices from the mixture of them.
  • Reproduced a RNN model (Source-Contrastive Estimation) basing on Monaural Audio Speaker Separation Using Source-Contrastive Estimation from 2017 IEEE International Workshop on Signal Processing Systems (SiPS)
  • Model exhibited excellent performance and project highly recognized by instructor.
Bruce (Zhi) Wen
Master Student in Computer Science