Compressive Detection of Highly Overlapped Spectra Using Walsh–Hadamard-Based Filter Functions

Abstract

In the chemometric context in which spectral loadings of the analytes are already known, spectral filter functions may be constructed which allow the scores of mixtures of analytes to be determined in on-the-fly fashion directly, by applying a compressive detection strategy. Rather than collecting the entire spectrum over the relevant region for the mixture, a filter function may be applied within the spectrometer itself so that only the scores are recorded. Consequently, compressive detection shrinks data sets tremendously. The Walsh functions, the binary basis used in Walsh-Hadamard transform spectroscopy, form a complete orthonormal set well suited to compressive detection. A method for constructing filter functions using binary fourfold linear combinations of Walsh functions is detailed using mathematics borrowed from genetic algorithm work, as a means of optimizing said functions for a specific set of analytes. These filter functions can be constructed to automatically strip the baseline from analysis. Monte Carlo simulations were performed with a mixture of four highly overlapped Raman loadings and with ten excitation-emission matrix loadings; both sets showed a very high degree of spectral overlap. Reasonable estimates of the true scores were obtained in both simulations using noisy data sets, proving the linearity of the method.

High-Sensitivity High-Speed Compressive Spectrometer for Raman Imaging

“…We benchmarked the spectrometer with two different types of compressive Raman microspectroscopy schemes: with (supervised) − ,, or without (unsupervised) , a priori knowledge about the principal components (or eigenvectors) of the hyperspectrum. In this section, we present compressive Raman results without a priori knowledge.…”

“…While compressive microspectroscopy − is a promising technology, it still lacks sensitivity for bioimaging when compared to sensitive CCD cameras. High-sensitivity imaging is of particular relevance in the sub-micrometer spatial scale, as it is the size range of many organelle in a cell, − , including lipid droplets, cell membranes, and membraneless organelles.…”

Recent Trends in Compressive Raman Spectroscopy Using DMD-Based Binary Detection

“…However, the efficiency of band-pass scanning (or notch-scanning) is quite low since most Raman photons are discarded. An advantageous alternative to notch-scanning may be to use Hadamard [ 27 ] filter functions to obtain full spectra at higher SNR [ 2 , 17 , 23 , 24 , 28 ]. The efficiency of Hadamard strategy comes from that half of the Raman photons is always detected by each Hadamard filter.…”

Portable Hadamard-Transform Raman Spectrometer: A Powerful Analytical Tool for Point-of-Care Testing

“…Besides, Hadamard-transform has the advantages of high energy input, single detector multichannel simultaneous detection (imaging), and high SNR. − Therefore, Hadamard-transform is suitable for weak spectra (image) analysis and POCT, applying it in the development of portable Raman spectrometer particularly. Recently, theoretical research on the application of Hadamard-transform to Raman spectrometer has been reported. , Besides, digital micromirror devices (DMDs) are programmable optical filters, which can be flipped at +12° and −12°, corresponding to the “on” and “off” states of the micromirror, respectively. Two states are associated with the “1” and “0” states of the signal binary.…”