Dr Hiroshi Sawada of NTT Communication Science Laboratories

Blind source separation (BSS) separates mixtures into individual sources with as little prior information as possible. Independent component analysis (ICA) is the primary method for BSS that estimates a linear transformation for separating a mixture into statistically independent components. ICA has many applications, including image and biomedical signal processing. For audio sources such as speech and music mixed in a real reverberant environment, we need to solve multiple ICA problems, each corresponding to a frequency of the audio signals. In addition, we need to roughly estimate individual sources to align the permutation indeterminacy of the ICA solutions. Nonnegative matrix factorization (NMF) provides a pathway to model each sound source by a low-rank approximation. NMF identifies frequent sound patterns in spectrograms. In this presentation, we combined NMF and ICA for simultaneous estimating of individual sources and linear transformations. We developed a sophisticated independent low-rank matrix analysis (ILRMA) method for audio BSS tasks. We will provide a live demonstration of BSS to separate two simultaneous speech sources recorded with a stereo IC recorder.

Biography:

Hiroshi Sawada received the B.E., M.E., and PhD degrees in information science from Kyoto University, Kyoto, Japan, in 1991, 1993, and 2001, respectively. In 1993, he joined NTT Corporation in Kyoto, Japan. He is currently a senior distinguished researcher at NTT Communication Science Laboratories, Kyoto, Japan. His research interests include statistical signal processing, audio source separation, array signal processing, machine learning, latent variable model, graph-based data structure, and computer architecture. He served as an Associate Editor for the IEEE Trans. Audio, Speech and Language Processing from 2006 to 2009, and as an Associate Editor of the IEEE Open Journal of Signal Processing from 2019 to 2022. He was a Member and an Associate Member of the Audio and Acoustic Signal Processing Technical Committee of the IEEE SP Society from 2006 to 2018. He received the Best Paper Award from the IEEE Circuit and System Society in 2000 and the Best Paper Award from the IEEE Signal Processing Society in 2014. He is an IEEE Signal Processing Society Distinguished Lecturer from 2022 to 2023. He is an IEEE Fellow, an IEICE Fellow, and a member of the Acoustical Society of Japan.