Independent component analysis: theory and applications
Independent component analysis: theory and applications
Independent Component Analysis: Principles and Practice
Independent Component Analysis: Principles and Practice
Nonholonomic Orthogonal Learning Algorithms for Blind Source Separation
Neural Computation
Evaluation of blind signal separation method using directivity pattern under reverberant conditions
ICASSP '00 Proceedings of the Acoustics, Speech, and Signal Processing, 2000. on IEEE International Conference - Volume 05
ICASSP '01 Proceedings of the Acoustics, Speech, and Signal Processing, 2001. on IEEE International Conference - Volume 05
Springer Handbook of Speech Processing
Springer Handbook of Speech Processing
Independent vector analysis: an extension of ICA to multivariate components
ICA'06 Proceedings of the 6th international conference on Independent Component Analysis and Blind Signal Separation
ICA'06 Proceedings of the 6th international conference on Independent Component Analysis and Blind Signal Separation
Multichannel signal separation: methods and analysis
IEEE Transactions on Signal Processing
Equivariant adaptive source separation
IEEE Transactions on Signal Processing
Blind Source Separation Exploiting Higher-Order Frequency Dependencies
IEEE Transactions on Audio, Speech, and Language Processing
Stability of independent vector analysis
Signal Processing
A matrix joint diagonalization approach for complex independent vector analysis
LVA/ICA'12 Proceedings of the 10th international conference on Latent Variable Analysis and Signal Separation
Hi-index | 0.00 |
Utilizing dependence over frequencies has shown significant excellence in tackling the frequency-domain blind source separation (BSS), which is also referred to as independent vector analysis (IVA). The IVA method then runs in offline batch processing, which is not well applicable to real-time systems. This paper proposes real-time BSS methods corresponding to that model. First, we derive online algorithms under some assumptions. Then, in order to improve the performance and convergence properties, a modified gradient with nonholonomic constraint and a gradient normalization method are proposed. The convergence speed is improved by the gradient normalization. The gradient with nonholonomic constraint shows better performances, although it has less computational complexity. In addition, the proposed method has a simpler structure than any other existing methods and runs in fully online mode. Thus, it requires sufficiently less computations and memories. Based on these benefits, the algorithm is implemented in a real-time embedded system. The experimental results confirm effectiveness of the proposed method with both simulated data and real recordings.