Band-Target Entropy Minimization (BTEM) Applied to Hyperspectral Raman Image Data

Abstract

Band-target entropy minimization (BTEM) has been applied to extraction of component spectra from hyperspectral Raman images. In this method singular value decomposition is used to calculate the eigenvectors of the spectroscopic image data set. Bands in non-noise eigenvectors that would normally be used for recovery of spectra are examined for localized spectral features. For a targeted (identified) band, information entropy minimization or a closely related algorithm is used to recover the spectrum containing this feature from the non-noise eigenvectors, plus the next 5-30 eigenvectors, in which noise predominates. Tests for which eigenvectors to include are described. The method is demonstrated on one synthesized Raman image data set and two bone tissue specimens. By inclusion of small amounts of signal that would be unused in other methods, BTEM enables the extraction of a larger number of component spectra than are otherwise obtainable. An improvement in signal/noise ratio of the recovered spectra is also obtained.

Recent advances in the development of Raman spectroscopy for deep non‐invasive medical diagnosis

“…at zero and non-zero spatial offsets) [1] scaled relative to each other to cancel the residual contribution from the surface layer present in the non-zero spectrum. For a multilayer sample or a sample with unknown number of layers processing would involve multivariate techniques [1,7], such as band targeted entropy minimization (BTEM) [8,9]. In both the cases the processing can be automated.…”

Estimating the number of pure chemical components in a mixture by maximum likelihood

“…Although factor analysis is a useful and popular technique, it frequently runs into difficulties with over-or underdetermination because it assumes that an appropriate number of eigenvectors has been selected [12,13]. In practice, the number of meaningful components (eigenvectors) in factor analysis is usually determined by the user through visual inspection, sometimes using prior knowledge on the chemical composition of the specimen.…”

“…In practice, the number of meaningful components (eigenvectors) in factor analysis is usually determined by the user through visual inspection, sometimes using prior knowledge on the chemical composition of the specimen. Alternatives to factor analysis include Simplisma and band-target entropy minimization (BTEM) [13]. BTEM allows the user to use forty or more additional eigenvalues and corresponding factors by performing exhaustive band targeting at different wavelengths.…”

“…BTEM allows the user to use forty or more additional eigenvalues and corresponding factors by performing exhaustive band targeting at different wavelengths. It has been shown to outperform Simplisma in terms of recovering minor components [13]. While the algorithm uses a large number of eigenvalues, the final decision on the number of extracted components still has to be made by the user and is based on visual inspection of the eigenvectors.…”

“…While the algorithm uses a large number of eigenvalues, the final decision on the number of extracted components still has to be made by the user and is based on visual inspection of the eigenvectors. It also places considerable demands on the user in terms of computational time and human interaction (manual rotation, exhaustive band targeting) [13]. This is especially undesirable if the user needs to analyze a large collection of Raman images or spectra.…”