Published paper on RSSMP with audio samples and Matlab code (available on FileExchange)
Published work on repetition-based audio source separation
Published BIRD denoising algorithm, outperforming state-of-the-art methods on synthetic and M/EEG data—even those with oracle noise knowledge
Released PyMP, an open-source Python package for greedy decompositions (including RSSMP), with documentation and reproducibility tutorials
Shared BIRD source code on GitHub (joint work with A. Gramfort)
Research Experience
Developed a variant of the Matching Pursuit (MP) algorithm during his PhD: Random Sequential Subdictionaries Matching Pursuit (RSSMP)
Applied RSSMP to enhance low-bitrate audio compression schemes
Proposed a source separation algorithm exploiting repeating musical patterns, combined with a sparse prior to form a structured sparsity model (MMV framework)
Developed the Blind Random Pursuit Denoising (BIRD) algorithm and its structured extension S-BIRD for denoising time series and M/EEG signals
Assisted his brother Yves, a renowned volcanologist, in field data analysis