Dynamic Independent Component/Vector Analysis: Time-Variant Linear Mixtures Separable by Time-Invariant Beamformers

@article{Koldovsk2021DynamicIC,
  title={Dynamic Independent Component/Vector Analysis: Time-Variant Linear Mixtures Separable by Time-Invariant Beamformers},
  author={Zbyněk Koldovsk{\'y} and V{\'a}clav Kautsk{\'y} and Petr Tichavsk{\'y} and Jaroslav Cmejla and Jiř{\'i} M{\'a}lek},
  journal={IEEE Transactions on Signal Processing},
  year={2021},
  volume={69},
  pages={2158-2173}
}
A novel extension of Independent Component and Independent Vector Analysis for blind extraction/separation of one or several sources from time-varying mixtures is proposed. The mixtures are assumed to be separable source-by-source in series or in parallel based on a recently proposed mixing model that allows for the movements of the desired source while the separating beamformer is time-invariant. The popular FastICA algorithm is extended for these mixtures in one-unit, symmetric and block… 

Figures and Tables from this paper

Algorithm for Independent Vector Extraction Based on Semi-Time-Variant Mixing Model
TLDR
A new algorithm for dynamic independent vector extraction is proposed based on the mixing model where mixing parameters related to the source-of-interest (SOI) are time-variant while the separating parameters areTime-invariant, which yields an algorithm that is significantly faster than gradient-based algorithms while different from fixed-point methods, which are even faster.
Blind Extraction of Moving Sources via Independent Component and Vector Analysis: Examples
TLDR
The recently proposed mixing model with constant separating vector (CSV) for Blind Source Extraction of moving sources using the FastDIVA algorithm, which is an extension of the famous FastICA and FastIVA for static mixtures shows their new potential and good performance in handling the blind moving source extraction problem.
Fasteriva: Update Rules for Independent Vector Analysis Based on Negentropy and the Majorize-Minimize Principle
TLDR
Fast converging update rules from a negentropy perspective are derived, based on the Majorize-Minimize (MM) principle and eigenvalue decomposition, which are shown to outperform competing state-of-the-art methods in terms of convergence speed at a comparable runtime.

References

SHOWING 1-10 OF 73 REFERENCES
Orthogonally constrained independent component extraction: Blind MPDR beamforming
TLDR
An algorithm that can be interpreted as a blind self-steering Minimum-Power Distortionless Response (MPDR) beamformer is proposed and compared with the popular Natural Gradient algorithm for general Independent Component Analysis.
Gradient Algorithms for Complex Non-Gaussian Independent Component/Vector Extraction, Question of Convergence
TLDR
The proposed modifications of the gradient methods, taking into account the dominance/weakness of the source, show improved global convergence.
Fast fixed-point independent vector analysis algorithms for convolutive blind source separation
TLDR
This work examines available contrasts for the new formulation that can solve the frequency-domain blind source separation problem and introduces a quadratic Taylor polynomial in the notations of complex variables which is very useful in directly applying Newton's method to a contrast function of complex-valued variables.
Fast Algorithm for Blind Independence-Based Extraction of a Moving Speaker
TLDR
The experiments corroborate the applicability of the CSV mixing model for the blind moving source extraction as well as the improved convergence of the proposed algorithms.
Performance Bound for Blind Extraction of Non-gaussian Complex-valued Vector Component from Gaussian Background
TLDR
A Cramér-Rao-Induced Bound (CRIB) for the achievable Interference-to-Signal Ratio (ISR) is derived and is compared with similar bounds for ICA, IVA, and Independent Component Extraction (ICE).
Independent Vector Analysis: Definition and Algorithms
We present a new approach to independent component analysis (ICA) by extending the formulation of univariate source signals to multivariate source signals. The new approach is termed independent
On the Performance of Orthogonal Source Separation Algorithms
Source separation consists in recovering a set of n independent signals from m n observed instantaneous mixtures of these signals, possibly corrupted by additive noise. Many source separation
A robust and precise method for solving the permutation problem of frequency-domain blind source separation
TLDR
By utilizing the harmonics of signals, the new method is robust even for low frequencies where DOA estimation is inaccurate, and provides an almost perfect solution to the permutation problem for a case where two sources were mixed in a room whose reverberation time was 300 ms.
Nonstationary Source Separation Using Sequential and Variational Bayesian Learning
TLDR
This paper presents an online learning algorithm for the Gaussian process (GP) and establishes a separation procedure in the presence of nonstationary and temporally correlated mixing coefficients and source signals and captures the evolved statistics from sequential signals according to online Bayesian learning.
Algorithm for Independent Vector Extraction Based on Semi-Time-Variant Mixing Model
TLDR
A new algorithm for dynamic independent vector extraction is proposed based on the mixing model where mixing parameters related to the source-of-interest (SOI) are time-variant while the separating parameters areTime-invariant, which yields an algorithm that is significantly faster than gradient-based algorithms while different from fixed-point methods, which are even faster.
...
1
2
3
4
5
...