In applications such as audio denoising, it is favorable to decompose a recording into its respective sources. Two commonly used methods are principal component analysis (PCA) and independent component analysis (ICA). However, such approaches perform poorly when no training data is given.
To overcome these issues, we propose another promising and effective algorithm, which is based on non-negative matrix factorization (NMF). The method works by decomposing the time-frequency domain of the signal into separated parts which correspond to respective sources.
The sparsity constraint is introduced to address the number of the bases. The sparse NMF is then proposed to find the appropriate number of bases during the factorization, which yields more convincing and accurate model of the original data.
To verify the advantage of the proposed method, we develop an acoustic fall detection system. For the simulated tasks, we find that sparse NMF based source separation is better than its NMF based counterpart in terms of approximation accuracy and detection performance. Overall, the result demonstrates that our proposed approach is a powerful separation method and motivates further work in source separation.
|School:||University of Missouri - Columbia|
|School Location:||United States -- Missouri|
|Source:||MAI 56/01M(E), Masters Abstracts International|
|Keywords:||Fall detection, NMF, Source separation|
Copyright in each Dissertation and Thesis is retained by the author. All Rights Reserved
The supplemental file or files you are about to download were provided to ProQuest by the author as part of a
dissertation or thesis. The supplemental files are provided "AS IS" without warranty. ProQuest is not responsible for the
content, format or impact on the supplemental file(s) on our system. in some cases, the file type may be unknown or
may be a .exe file. We recommend caution as you open such files.
Copyright of the original materials contained in the supplemental file is retained by the author and your access to the
supplemental files is subject to the ProQuest Terms and Conditions of use.
Depending on the size of the file(s) you are downloading, the system may take some time to download them. Please be