Blind spectral unmixing in terahertz domain using nonnegative matrix factorization

Xian Li; Ping J. Huang; Ye H. Ma; Di B. Hou; Guangxin Zhang

doi:10.1117/12.2212263

5 November 2015 Blind spectral unmixing in terahertz domain using nonnegative matrix factorization

Xian Li, Ping J. Huang, Ye H. Ma, Di B. Hou, Guangxin Zhang

Proceedings Volume 9795, Selected Papers of the Photoelectronic Technology Committee Conferences held June–July 2015; 97953E (2015) https://doi.org/10.1117/12.2212263
Event: Selected Proceedings of the Photoelectronic Technology Committee Conferences held June-July 2015, 2015, Hefei, Suzhou, and Harbin, China

Abstract

Recovering component spectra from terahertz measurements of unknown mixtures has been studied in this paper using nonnegative matrix factorization (NMF). NMF mathematically decomposes the spectra data into two nonnegative matrixes which describe the component spectra and the corresponding fractional abundance. Two basic algorithms in the class of this method, NMF and NMF with smoothness constraint (cNMF), were adopted to resolve the terahertz absorption spectra matrix obtained from a ternary mixture with varying compositions of Nitrofurantoin, L-Leucine and D-Tyrosine. The quality of the decomposition results was evaluated. The performance of the two algorithms on extracting component terahertz spectra was compared. The optimal result reached by cNMF in this study implies the capability of the NMF method for blind terahertz spectral unmixing. The attempt made in our work helps to further investigate unknown mixtures by terahertz spectroscopy.

Citation Download Citation

Xian Li, Ping J. Huang, Ye H. Ma, Di B. Hou, and Guangxin Zhang "Blind spectral unmixing in terahertz domain using nonnegative matrix factorization", Proc. SPIE 9795, Selected Papers of the Photoelectronic Technology Committee Conferences held June–July 2015, 97953E (5 November 2015); https://doi.org/10.1117/12.2212263

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

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.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available