A method for target speech enhancement based on degenerate unmixing and estimating technique (DUET) has
been described. To avoid the requirements of the DUET which need to know the number of sources in advance
and to estimate the attenuation and delay parameters for all sources, the method assumes that extraction of only
one target signal is required, which is often plausible in real-world applications such as speech enhancement. The
method can efficiently recover the target speech with fast convergence by estimating the parameters for the target
source only. In addition, it does not need to know the number of sources in advance. In order to accomplish robust
speech recognition, we propose an algorithm which employs the cluster-based missing feature reconstruction
technique based on log-spectral features of enhanced speech in the process of extracting mel-frequency cepstral
coefficients (MFCCs). The algorithm estimates missing time-frequency regions by computing the signal-to-noise
ratios (SNRs) from the log-spectral features of the enhanced speech and observed noisy speech and by finding time-frequency segments which have the SNRs smaller than a threshold. The missing time-frequency regions are filled by using bounded estimation based on the log-spectral features that are considered to be reliable and on the knowledge of the log-spectral feature cluster to which the incoming target speech is assumed to belong. Then, the log-spectral features are transformed into cepstral features in the usual fashion of extracting MFCCs. Experimental results show that the proposed algorithm significantly improves recognition performance in noisy environments.
Independent component analysis (ICA) for acoustic mixtures has been a challenging problem due to very complex
reverberation involved in real-world mixing environments. In an effort to overcome disadvantages of the
conventional time domain and frequency domain approaches, this paper describes filterbank-based independent
component analysis for acoustic mixtures. In this approach, input signals are split into subband signals and
decimated. A simplified network performs ICA on the decimated signals, and finally independent components
are synthesized. First, a uniform filterbank is employed in the approach for basic and simple derivation and implementation.
The uniform-filterbank-based approach achieves better separation performance than the frequency
domain approach and gives faster convergence speed with less computational complexity than the time domain
approach. Since most of natural signals have exponentially or more steeply decreasing energy as the frequency
increases, the spectral characteristics of natural signals introduce a Bark-scale filterbank which divides low frequency
region minutely and high frequency region widely. The Bark-scale-filterbank-based approach shows faster
convergence speed than the uniform-filterbank-based one because it has more whitened inputs in low frequency
subbands. It also improves separation performance as it has enough data to train adaptive parameters exactly
in high frequency subbands.
Conference Committee Involvement (4)
Independent Component Analyses, Compressive Sampling, Large Data Analyses (LDA), Neural Networks, Biosystems, and Nanoengineering XIII
23 April 2015 | Baltimore, MD, United States
Independent Component Analyses, Compressive Sampling, Wavelets, Neural Net, Biosystems, and Nanoengineering XII
7 May 2014 | Baltimore, MD, United States
Independent Component Analyses, Compressive Sampling, Wavelets, Neural Net, Biosystems, and Nanoengineering XI
1 May 2013 | Baltimore, Maryland, United States
Independent Component Analyses, Compressive Sampling, Wavelets, Neural Net, Biosystems, and Nanoengineering X
25 April 2012 | Baltimore, Maryland, United States
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.