Single channel speech enhancement using Bayesian NMF with recursive temporal updates of prior distributions

Mohammadiha, Nasser; Taghia, Jalil; Leijon, Arne

doi:10.1109/icassp.2012.6288933

Cited by 35 publications

(41 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A linear minimum mean square error (LMMSE) estimator was proposed in [10] for speech enhancement where the temporal dynamics were used in filter construction. In [11], an on-line speech enhancement algorithm was proposed in which temporal aspects of the data were used to obtain informative prior distributions to be applied in a Bayesian NMF framework.…”

Section: Introductionmentioning

confidence: 99%

Prediction based filtering and smoothing to exploit temporal dependencies in NMF

Mohammadiha

Smaragdis

Leijon

2013

2013 IEEE International Conference on Acoustics, Speech and Signal Processing

Self Cite

View full text Add to dashboard Cite

Nonnegative matrix factorization is an appealing technique for many audio applications. However, in it's basic form it does not use temporal structure, which is an important source of information in speech processing. In this paper, we propose NMF-based filtering and smoothing algorithms that are related to Kalman filtering and smoothing. While our prediction step is similar to that of Kalman filtering, we develop a multiplicative update step which is more convenient for nonnegative data analysis and in line with existing NMF literature. The proposed smoothing approach introduces an unavoidable processing delay, but the filtering algorithm does not and can be readily used for on-line applications. Our experiments using the proposed algorithms show a significant improvement over the baseline NMF approaches. In the case of speech denoising with factory noise at 0 dB input SNR, the smoothing algorithm outperforms NMF with 3.2 dB in SDR and around 0.5 MOS in PESQ, likewise source separation experiments result in improved performance due to taking advantage of the temporal regularities in speech.

show abstract

Section: Introductionmentioning

confidence: 99%

Prediction based filtering and smoothing to exploit temporal dependencies in NMF

Mohammadiha

Smaragdis

Leijon

2013

2013 IEEE International Conference on Acoustics, Speech and Signal Processing

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, for the noise reduction even if a universal speaker-independent basis matrix of speech is learned a good enhancement can be achieved [4]. In some cases when the interfering noise exhibits speech-like properties, e.g.…”

Section: Introductionmentioning

confidence: 99%

“…In speech processing, Y is usually the spectrogram of the speech signal. NMF has been recently used to estimate the clean speech from a noisy observation [3,4,6,10]. A main focus of most of these approaches is to use the temporal dynamics in an NMF-based speech denoising or separation algorithm [4,6,11].…”

Section: Introductionmentioning

confidence: 99%

“…Methods proposed in [4,6,12] require to know the noise type in advance to use noise-dependent basis matrices to enhance the noisy signal. For some applications where user is allowed to choose a scenario from a given set, this assumption can be feasible.…”

Section: Introductionmentioning

confidence: 99%

“…Then, assuming that the noise type is known, the noise reduction task is performed. Some examples of this class of algorithms include hidden Markov model (HMM) based methods [2] and NMF-based techniques [3,4].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations