BIRD: Python implementation of Blind Random Pursuit Denoising
By M. Moussalam

Pure Python implementation of the BIRD algorithm. BIRD is a denoising algorithm using randomized greedy pursuits. It works for single signals and supports (structured)-sparsity for multivariate signals, e.g. multichannel array data.

The algorithm presented here are described in:

Blind Denoising with Random Greedy Pursuits. Moussallam, M., Gramfort, A., Daudet, L., & Richard, G. (2014). IEEE Signal Processing Letters, 21(11), 1341-1345 Preprint available at:

Obtain BIRD: on GitHub