 |
|---|
| Illustration of an example for an extedned SeDJoCo transformation. |
The extended “Sequentially Drilled” Joint Congruence (SeDJoCo) transformation is a special joint matrix transformation, reminiscent of (but different from) classical joint diagonalization. Interestingly, it turns out that the Maximum Likelihood (ML) solution for the Independent Vector Analysis (IVA) problem with a Gaussian model takes the form of an extended SeDJoCo problem.
This package contains five files: a (readme) detailed instruction file, two functions for an iterative solution of extended SeDJoCo—iterative relaxations and Newton’s method—and two scripts which demonstrate their operation. The first script solves a generic problem, while the second demonstrates how it is used for the computation of the ML solution of a Gaussian IVA problem. For more details, see [1].
To download the Matlab package, click here.
Related Papers
[1] Weiss, A., Yeredor, A., Cheema, S. A. and Haardt, M., “The Extended “Sequentially Drilled” Joint Congruence Transformation and its Application in Gaussian Independent Vector Analysis”, IEEE Trans. on Signal Processing, vol. 65, no. 23, pp. 1-13, Dec. 2017. arXiv
[2] Weiss, A., Yeredor, A., Cheema, S. A. and Haardt, M., “Performance Analysis of the Gaussian Quasi-Maximum Likelihood Approach for Independent Vector Analysis”, IEEE Trans. on Signal Processing, vol. 66, no. 19, pp. 5000-5013, Sept. 2018. arXiv
[3] Cheng, Y., Yeredor, A., Weiss, A. and Haardt, M., “Extension of the “Sequentially Drilled” Joint Congruence Transformation (SeDJoCo) Problem”, in Proc. of IEEE 6th Int. Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP), pp. 185–188, Dec. 2015.