Blind Source Separation On Biomedical Field By Using Nonnegative Matrix Factorization
The study of separating heart from lung sound has been investigated and researched for years. However, a novel approach based on nonnegative matrix factorization (NMF) as a skill of blind source separation (BSS) that utilized in biomedical field is fresh presented. Lung sound gives beneficial info...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | en |
| Published: |
Asian Research Publishing Network (ARPN)
2016
|
| Subjects: | |
| Online Access: | http://eprints.utem.edu.my/id/eprint/17101/1/Blind%20Source%20Separation%20On%20Biomedical%20Field%20By%20Using%20Nonnegative%20Matrix%20Factorization.pdf http://eprints.utem.edu.my/id/eprint/17101/ http://www.arpnjournals.org/jeas/research_papers/rp_2016/jeas_0716_4584.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The study of separating heart from lung sound has been investigated and researched for years. However, a novel
approach based on nonnegative matrix factorization (NMF) as a skill of blind source separation (BSS) that utilized in
biomedical field is fresh presented. Lung sound gives beneficial information regarding lung status through respiratory analysis. However, interrupt of heart sound is the obstacle from taking precise and exact information during respiratory analysis. Thus, separation heart sound from lung sound is a way to overcome this issue in order to determine the accuracy of respiratory analysis. This paper proposes factorizations approach that concern on the 2 dimensional which is combination of frequency domain and time domain or well known as NMF2D. The proposed method is developed under the divergence of Least Square Error and Kullback-Leibler and it demonstrates from a single channel source. In this paper, we will forms a multivariate data and it will proceed for dimension reduction by log frequency domain. Experimental tests and comparisons will be made via different divergence to verify and evaluate efficiency of the proposed method in term performance measurement. |
|---|