Spectral unmixing combined with Raman imaging, a preferable analytic technique for molecule visualization

ABSTRACT

This article reviews spectral unmixing used as a new method for Raman imaging. Raman spectroscopy explores the vibrational information about specific chemical bonds in molecules, which can be used for label-free molecular visualization. However, chemical bonds are usually shared among different molecules, which results in closed or mixed Raman peaks of many molecules. Therefore, the acquired spectra cannot be directly used to reconstruct the Raman images, as pure component spectra are hidden under the acquired spectra. Spectral unmixing is an effective method to provide a meaningful spectrum of each component with no priori spectral information and to also reconstruct the compositional distribution images. This article summarizes some representative spectral unmixing approaches used for Raman imaging and many related researches. This review strives to introduce the combination of spectral unmixing and Raman imaging as an efficient analytic technique to characterize various constituents and make subtle understanding of those complex structures.

KEYWORDS:

• Biological visualization
• biomedical diagnosis
• compositional distribution analysis
• Raman imaging
• spectral unmixing

Funding

This research was supported by the National Natural Science Foundation of China (61201073), the Fundamental Research Funds for the Central Universities and Zhejiang province public technology research program (2014C32091), the Sub-project under National Science and Technology Support Program (2014BAD06B06), and Scientific research projects of the Education Department of Zhejiang Province (Y201225966).