-
- Rintaro Ikeshita
- NTT Corporation,NTT Communication Science Laboratories,Kyoto,Japan
-
- Nobutaka Ito
- NTT Corporation,NTT Communication Science Laboratories,Kyoto,Japan
-
- Tomohiro Nakatani
- NTT Corporation,NTT Communication Science Laboratories,Kyoto,Japan
-
- Hiroshi Sawada
- NTT Corporation,NTT Communication Science Laboratories,Kyoto,Japan
書誌事項
- 公開日
- 2019-10
- 権利情報
-
- https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html
- https://doi.org/10.15223/policy-029
- https://doi.org/10.15223/policy-037
- DOI
-
- 10.1109/waspaa.2019.8937171
- 公開者
- IEEE
説明
This paper addresses the determined convolutive blind source separation (BSS) problem. The state-of-the-art independent low-rank matrix analysis (ILRMA), unifying independent component analysis (ICA) and nonnegative matrix factorization, has the disadvantage of ignoring inter-frame and inter-frequency spectral correlation of source signals. We here propose a new BSS method that estimates a linear transformation for spectral decorrelation and performs ILRMA in the transformed domain. A newly introduced optimization problem is an extension of that for ICA based on maximum likelihood. For this problem, we provide a necessary and sufficient condition for the existence of optimal solutions, and develop algorithms based on block coordinate descent methods with closed-form solutions. Experimental results show the improved separation performance of the proposed method compared to ILRMA.
収録刊行物
-
- 2019 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA)
-
2019 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA) 288-292, 2019-10
IEEE