Fast unmixing of multispectral optoacoustic data with vertex component analysis

Dean-Ben XL, Deliolanis NC, Ntziachristos V, Razansky D (2014)

Publication Type: Journal article

Publication year: 2014

Journal

Optics and Lasers in Engineering Elsevier

Book Volume: 58

Pages Range: 119-125

DOI: 10.1016/j.optlaseng.2014.01.027

Abstract

Multispectral optoacoustic tomography enhances the performance of single-wavelength imaging in terms of sensitivity and selectivity in the measurement of the biodistribution of specific chromophores, thus enabling functional and molecular imaging applications. Spectral unmixing algorithms are used to decompose multi-spectral optoacoustic data into a set of images representing distribution of each individual chromophoric component while the particular algorithm employed determines the sensitivity and speed of data visualization. Here we suggest using vertex component analysis (VCA), a method with demonstrated good performance in hyperspectral imaging, as a fast blind unmixing algorithm for multispectral optoacoustic tomography. The performance of the method is subsequently compared with a previously reported blind unmixing procedure in optoacoustic tomography based on a combination of principal component analysis (PCA) and independent component analysis (ICA). As in most practical cases the absorption spectrum of the imaged chromophores and contrast agents are known or can be determined using e.g. a spectrophotometer, we further investigate the so-called semi-blind approach, in which the a priori known spectral profiles are included in a modified version of the algorithm termed constrained VCA. The performance of this approach is also analysed in numerical simulations and experimental measurements. It has been determined that, while the standard version of the VCA algorithm can attain similar sensitivity to the PCA-ICA approach and have a robust and faster performance, using the a priori measured spectral information within the constrained VCA does not generally render improvements in detection sensitivity in experimental optoacoustic measurements. © 2014 Elsevier Ltd.

Involved external institutions

Technische Universität München (TUM) DE Germany (DE) Fraunhofer-Institut für Produktionstechnik und Automatisierung (IPA) DE Germany (DE)

How to cite

APA:

Dean-Ben, X.L., Deliolanis, N.C., Ntziachristos, V., & Razansky, D. (2014). Fast unmixing of multispectral optoacoustic data with vertex component analysis. Optics and Lasers in Engineering, 58, 119-125. https://doi.org/10.1016/j.optlaseng.2014.01.027

MLA:

Dean-Ben, X. Luis, et al. "Fast unmixing of multispectral optoacoustic data with vertex component analysis." Optics and Lasers in Engineering 58 (2014): 119-125.

BibTeX: Download