Fast spectrum extraction from SRS spectral images by independent component analysis with image compression and post sampling

摘要

Stimulated Raman scattering (SRS) spectral imaging combined with multivariate analysis such as independent component analysis (ICA) allows rapid imaging of several constituents with spectral information. However, ICA takes several tens of seconds and it hinders real-time analysis. Furthermore, spectra given by ICA are not always identical to the original spectral images. In this paper, I show that ICA can be very quick by compressing the images. Furthermore, I propose a technique for extracting spectral information based on the post sampling at the locations where the ICA intensity is high. I demonstrate the proposed technique for the imaging of polymer beads and tissue samples.