I
Ina Kodrasi
Researcher at Idiap Research Institute
Publications - 53
Citations - 606
Ina Kodrasi is an academic researcher from Idiap Research Institute. The author has contributed to research in topics: Wiener filter & Speech enhancement. The author has an hindex of 12, co-authored 48 publications receiving 452 citations. Previous affiliations of Ina Kodrasi include University of Oldenburg & Jacobs University Bremen.
Papers
More filters
Journal ArticleDOI
Regularization for Partial Multichannel Equalization for Speech Dereverberation
TL;DR: A partial multichannel equalization technique based on MINT (P-MINT) is proposed which aims to shorten the RIR and an automatic non-intrusive procedure for determining the regularization parameter based on the L-curve is introduced.
Journal ArticleDOI
Combination of MVDR beamforming and single-channel spectral processing for enhancing noisy and reverberant speech
Benjamin Cauchi,Ina Kodrasi,Robert Rehr,Stephan Gerlach,Ante Jukic,Timo Gerkmann,Simon Doclo,Stefan Goetze +7 more
TL;DR: Experimental results show that the proposed system is effective in suppressing both reverberation and noise while improving the speech quality, and the achieved improvements are particularly significant in conditions with high reverberation times.
Journal ArticleDOI
Spectro-Temporal Sparsity Characterization for Dysarthric Speech Detection
Ina Kodrasi,Hervé Bourlard +1 more
TL;DR: A numerical analysis of the suitability of different non-parametric and parametric measures for sparsity characterization shows that kurtosis, the Gini index, and the parametric sparsity measures are advantageous sparsity Measures, whereas the $l_1$-norm and entropy measures fail to robustly characterize the temporal sparsity of signals with a different number of time frames.
Proceedings ArticleDOI
Subjective speech quality and speech intelligibility evaluation of single-channel dereverberation algorithms
Anna Warzybok,Ina Kodrasi,Jan Ole Jungmann,Emanuel A. P. Habets,Timo Gerkmann,Alfred Mertins,Simon Doclo,Birger Kollmeier,Stefan Goetze +8 more
TL;DR: Six different single-channel dereverberation algorithms are evaluated subjectively in terms of speech intelligibility and speech quality and the best performance was observed for the regularized spectral inverse approach with pre-echo removal.
Journal ArticleDOI
Joint dereverberation and noise reduction based on acoustic multi-channel equalization
Ina Kodrasi,Simon Doclo +1 more
TL;DR: Two time-domain techniques aiming at joint dereverberation and noise reduction based on acoustic multi-channel equalization based on multiple-input/output inverse theorem are proposed and it is shown that the MWF-D NR technique yields a significantly better noise reduction performance than the RPM-DNR technique at the expense of a worse derever beration performance.