Signal Processing and Speech Communication Laboratory
homestudent projects › Realtime Model-based Single-Channel Speech Separation

Realtime Model-based Single-Channel Speech Separation

Status
Open
Type
Master/Diploma Thesis
Seminar Type
- None -
Announcement date
06 Mar 2013
Mentors
Research Areas

Short Description

While many advances have been made in single-channel speech separation problem [1-5], the current state-of-the-art methods lag far behind to be real-time implemented. Current model-driven methods are highly computationally complex and still lag far behind in being implemented in real-time.

Your Tasks

  • Balance a trade-off between the separation accuracy and speed of current method
  • Preparing a realtime demonstration for single-channel source separation

Your Profile

The candidate should be interested in speech signal processing, Matlab and C programming.

References

[1] P. Mowlaee, “New strategies for single-channel speech separation”, PhD thesis, Department of Electronic systems Aalborg University, 2010.
[2] P. Mowlaee, R. Saeidi, Z.-H, Tan, M. G. Christensen, T. Kinnunen, P. Fränti, S. H. Jensen, “A Joint Approach for single-channel Speaker Identification and Speech Separation”, IEEE Trans. on Audio, Speech, and Language Process., vol. 20, no. 9, pp. 2586-2601, 2012.
[3] P. Mowlaee, M. G. Christensen, S. H. Jensen, “New Results on Single-Channel Speech Separation Using Sinusoidal Modeling”, Audio, Speech, and Language Processing, IEEE Transactions on , vol.19, no.5, pp.1265-1277, 2011.
[4] Stark, M., M. Wohlmayr, and F. Pernkopf, “Source-Filter based Single Channel Speech Separation using Pitch Information”, IEEE Transactions on Audio, Speech and Language Processing, vol. 19, issue 2, pp. 242 - 255, Feb., 2011
[5] M. Cooke, J. R. Hershey, and S. J. Rennie, “Monaural speech separation and recognition challenge”, Elsevier Comput. Speech Lang., vol. 24, no. 1, pp. 1-15, 2010