Convolutive BSS of Short Mixtures by ICA Recursively Regularized Across Frequencies

Nesta, Francesco; Svaizer, Piergiorgio; Omologo, Maurizio

doi:10.1109/tasl.2010.2053027

Cited by 72 publications

(51 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The results support the emergence of source separation systems exploiting advanced source models accounting for the source spectra in the case of audio source separation [22,23,24,30] or for signaling pathway information in the case of biomedical source separation [34]. Nevertheless, more conventional methods based on frequency-domain ICA or SCA still perform best on live audio recordings of many sources and/or background noise [25,27,28,29].…”

Section: Remaining Challengessupporting

confidence: 54%

“…• Similar performance is achieved over 2-channel noiseless mixtures of 2 sources, again by means of frequency-domain ICA [29]. Note that the considered 2-channel 2-source mixtures were either short or dynamic, which shows that frequency-domain ICA can efficiently adapt to such situations [29].…”

Section: Current Performance On the Other Audio Datasetsmentioning

confidence: 52%

“…Note that the considered 2-channel 2-source mixtures were either short or dynamic, which shows that frequency-domain ICA can efficiently adapt to such situations [29]. These methods result in significant filtering distortion of the source signals, however, as indicated by the lower SAR.…”

Section: Current Performance On the Other Audio Datasetsmentioning

confidence: 99%

See 2 more Smart Citations

The signal separation evaluation campaign (2007–2010): Achievements and remaining challenges

et al. 2012

View full text Add to dashboard Cite

We present the outcomes of three recent evaluation campaigns in the field of audio and biomedical source separation. These campaigns have witnessed a boom in the range of applications of source separation systems in the last few years, as shown by the increasing number of datasets from 1 to 9 and the increasing number of submissions from 15 to 34. We first discuss their impact on the definition of a reference evaluation methodology, together with shared datasets and software. We then present the key results obtained over almost all datasets. We conclude by proposing directions for future research and evaluation, based in particular on the ideas raised during the related panel discussion at the Ninth International Conference on Latent Variable Analysis and Signal Separation (LVA/ICA 2010).

show abstract

Section: Remaining Challengessupporting

confidence: 54%

Section: Current Performance On the Other Audio Datasetsmentioning

confidence: 52%

See 1 more Smart Citation

The signal separation evaluation campaign (2007–2010): Achievements and remaining challenges

et al. 2012

View full text Add to dashboard Cite

show abstract

“…Therefore, the first step in information processing of speech signal is usually speech separation from the contaminated inputs in order to acquire a good front-end model. Up to now, many approaches have been proposed for speech separation, including independent component analysis (ICA) [1], non-negative matrix factorization (NMF) [2], subspace decomposition algorithm [3], and tools in computational auditory scene analysis [4] etc. Those approaches can only obtain good performance on speech separation when the target speech and interference signals satisfy certain constraints.…”

Section: Introductionmentioning

confidence: 99%

Speech Separation based on Deep Belief Network

Wu¹,

Zhang²,

Zhang³

et al. 2015

Proceedings of the 2015 International Industrial Informatics and Computer Engineering Conference

View full text Add to dashboard Cite

Abstract. Thanks to its hierarchical and generative nature, Deep Belief Network (DBN) is effective to feature representation and extraction in signal processing. In this paper, DBN is investigated and implemented to monaural speech separation. Firstly, two separate DBNs are trained to extract features from mixed noisy signals and target clean speech respectively. Subsequently, the two types of extracted features are associated together by training a BP neural network to obtain a mapping from the features of mixed signals to the features of target speech. Finally, by performing DBN and the above mapping neural network, target speech can be estimated from the input mixed signals. Experiments are conducted on different kinds of mixed signals including female/male speech mixtures, human-speech/Gaussian-noise audio mixtures, and human-speech/music audio mixtures. The PESQ scores of the extracted speech are 3.32, 2.59, and 3.42 respectively, which illustrates that the model performs well on speech separation tasks, especially on the mixed signals where the inference signals have obvious spectral structures.

show abstract

“…In the papers, many modified ICA methods have been proposed, such as infomax ICA [2,3], JADE [4], SOBI [5] , fast fixed point algorithm [6] , H-J [7], etc. Early, ICA resulted from the classic blind source separation (BSS) problem of a cocktailparty with less priori knowledge or even nothing [8]. ICA defines a generative model for the observed multivariate data from a large database of samples.…”

Section: Introductionmentioning

confidence: 99%