Daichi Kitamura

Determined blind source separation unifying independent vector analysis and nonnegative matrix factorization

This paper addresses the determined blind source separation problem and proposes a new effective method unifying independent vector analysis (IVA) and nonnegative matrix factorization (NMF). IVA is a state-of-the-art technique that utilizes the statistical independence between sources in a mixture signal, and an efficient optimization scheme has been proposed for IVA. However, since the source model in IVA is based on a spherical multivariate distribution, IVA cannot utilize specific spectral structures such as the harmonic structures of pitched instrumental sounds. To solve this problem, we introduce NMF decomposition as the source model in IVA to capture the spectral structures. The formulation of the proposed method is derived from conventional multichannel NMF (MNMF), which reveals the relationship between MNMF and IVA. The proposed method can be optimized by the update rules of IVA and single-channel NMF. Experimental results show the efficacy of the proposed method compared with IVA and MNMF in terms of separation accuracy and convergence speed.

Efficient multichannel nonnegative matrix factorization exploiting rank-1 spatial model

This paper proposes a new efficient multichannel nonnegative matrix factorization (NMF) method. Recently, multichannel NMF (MNMF) has been proposed as a means of solving the blind source separation problem. This method estimates a mixing system of sources and attempts to separate them in a blind fashion. However, this method is strongly dependent on its initial values because there are no constraints in the spatial models. To solve this problem, we introduce a rank-1 spatial model into MNMF. The proposed method estimates a demixing matrix while representing sources using NMF bases and can be optimized by the update rules of independent vector analysis and conventional single-channel NMF. Experimental results show the efficacy of the proposed method in terms of robustness and convergence speed.

Multichannel signal separation combining directional clustering and nonnegative matrix factorization with spectrogram restoration

In this paper, to address problems in multichannel music signal separation, we propose a new hybrid method that combines directional clustering and advanced nonnegative matrix factorization (NMF). The aims of multichannel music signal separation technology is to extract a specific target signal from observed multichannel signals that contain multiple instrumental sounds. In previous studies, various methods using NMF have been proposed, but many problems remain including poor separation accuracy and lack of robustness. To solve these problems, we propose a new supervised NMF (SNMF) with spectrogram restoration and a hybrid method that concatenates the proposed SNMF after directional clustering. Via the extrapolation of supervised spectral bases, the proposed SNMF attempts both target signal separation and reconstruction of the lost target components, which are generated by preceding directional clustering. In addition, we experimentally reveal the trade-off between separation and extrapolation abilities and propose a new scheme for adaptive divergence, where the optimal divergence can be automatically changed in each time frame according to the local spatial conditions. The results of an evaluation experiment show that our proposed hybrid method outperforms the conventional music signal separation methods.

Relaxation of rank-1 spatial constraint in overdetermined blind source separation

In this paper, we propose a new algorithm for overdetermined blind source separation (BSS), which enables us to achieve good separation performance even for signals recorded in a reverberant environment. The proposed algorithm utilizes ex tra observations (channels) in overdetermined BSS to esti mate both direct and reverberant components of each source. This approach can relax the rank-1 spatial constraint, which corresponds to the assumption of a linear time-invariant mixing system. To confirm the efficacy of the proposed algorithm, we apply the relaxation of the rank-1 spatial constraint to con ventional BSS techniques. The experimental results show that the proposed algorithm can avoid the degradation of separation performance for reverberant signals in some cases.

Music signal separation by supervised nonnegative matrix factorization with basis deformation

In this paper, we address a music signal separation problem, and propose a new supervised algorithm for real instrumental signal separation employing a deformable capability for a spectral supervision trained in advance. Nonnegative matrix factorization (NMF) is one of the techniques used for the separation of an audio mixture that consists of multiple instrumental sources. Conventional supervised NMF has the critical problem that a mismatch between the bases trained in advance and the target real sound reduces the accuracy of separation. To solve this problem, we propose a new advanced supervised NMF that employs a deformable capability for the trained bases and penalty terms for making the bases fit into the target sound. The results of the experiment using real instruments show that the proposed method significantly improves the accuracy of separation compared with the conventional method.

Superresolution-based stereo signal separation via supervised nonnegative matrix factorization

In this paper, we address a stereo signal separation problem and propose a new method utilizing both directional clustering and superresolution-based supervised nonnegative matrix factorization (NMF) via spectrogram extrapolation using supervised bases. In previous studies, a hybrid method concatenating supervised NMF after directional clustering was proposed as for multichannel signal separation. However, this hybrid method has a problem that the extracted signal suffers from considerable spectral distortion because directional clustering yields spectral chasms. To solve this problem, we propose a new supervised NMF algorithm that regards the spectral chasms as unseen observations and reconstructs the target source components via spectrogram extrapolation using supervised bases. Our experimental results show that the proposed method outperforms several conventional methods and that the distortion of the extracted signal can be mitigated by superresolution efficacy.

Blind source separation based on independent low-rank matrix analysis with sparse regularization for time-series activity

In this paper, we propose a new blind source separation (BSS) method based on independent low-rank matrix analysis (ILRMA) with novel sparse regularization. ILRMA is a recently proposed BSS algorithm that simultaneously estimates a demixing matrix and source spectrogram models based on nonnegative matrix factorization (NMF). To improve the separation accuracy and stability, an additional constraint such as sparseness is needed but there have been no studies on this so far. In this study, we introduce an a priori statistical model for time-series amplitudes of source spectrograms, employing a new frequency-wise sparse regularization using estimates from the Bayesian postfilter to enhance the modeling accuracy. This regularization results in a bilevel optimization problem that consists of the estimation of a sparsity-emphasized source model using NMF and the separation of sources by ILRMA. In this paper, we present two approximated optimization schemes and their combination for performing regularized ILRMA. The efficacy of the proposed method is confirmed in a BSS experiment.

Robust music signal separation based on supervised nonnegative matrix factorization with prevention of basis sharing

In this paper, we address a monaural source separation problem and propose a new penalized supervised nonnegative matrix factorization (SNMF). Conventional SNMF often degrades the separation performance owing to the basis-sharing problem between supervised bases and nontarget bases. To solve this problem, we employ two types of penalty term based on orthogonality and divergence maximization in the cost function to force the nontarget bases to become as different as possible from the supervised bases. From the experimental results, it can be confirmed that the proposed method prevents the simultaneous generation of similar spectral patterns in the supervised bases and other bases, and increases the separation performance compared with the conventional method.

Low latency and high quality two-stage human-voice-enhancement system for a hose-shaped rescue robot

2-1-2 Hitotsubashi, Chiyoda-ku, Tokyo 101-8430, Japan ∗4Graduate School of Systems and Information Engineering, Tsukuba University 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8577, Japan ∗5Department of Informatics, School of Multidisciplinary Sciences, SOKENDAI 2-1-2 Hitotsubashi, Chiyoda-ku, Tokyo 101-8430, Japan ∗6Graduate School of Information Science, Tohoku University 6-6-01 Aramaki Aza Aoba, Aoba-ku, Sendai 980-8579, Japan ∗7Graduate Program for Embodiment Informatics, Waseda University 2-4-12 Okubo, Shinjuku, Tokyo 169-0072, Japan

Music signal separation based on Bayesian spectral amplitude estimator with automatic target prior adaptation

In this paper, we propose a new approach for addressing music signal separation based on the generalized Bayesian estimator with automatic prior adaptation. This method consists of three parts, namely, the generalized MMSE-STSA estimator with a flexible target signal prior, the NMF-based dynamic interference spectrogram estimator, and closed-form parameter estimation for the statistical model of the target signal based on higher-order statistics. The statistical model parameter of the hidden target signal can be detected automatically for optimal Bayesian estimation with online target-signal prior adaptation. Our experimental evaluation can show the efficacy of the proposed method.