Estimating and correcting interference fringes in infrared spectra in infrared hyperspectral imaging

Abstract

Short-term acclimation response of individual cells of Thalassiosira weissflogii was monitored by Synchrotron FTIR imaging over the span of 75 minutes. The cells, collected from batch cultures, were maintained in a constant flow of medium, at an irradiance of 120 μmol m-2 s-1 and at 20 °C. Multiple internal reflections due to the micro fluidic channel were modeled, and showed that fringes are additive sinusoids to the pure absorption of the other components of the system. Preprocessing of the hyperspectral cube (x, y, Abs(λ)) included removing spectral fringe using an EMSC approach. Principal component analysis of the time series of hyperspectral cubes showed macromolecular pool variations (carbohydrates, lipids and DNA/RNA) of less than 2% after fringe correction.

An automated approach for fringe frequency estimation and removal in infrared spectroscopy and hyperspectral imaging of biological samples

“…This strategy is therefore proposed as a step in the Fringe EMSC algorithm. In the Fringe EMSC algorithm proposed by Azarfar et al [16], they suggest to use multiple iterations. However, by selecting frequencies from multiple maxima in the Fourier domain, we experience that an iterative approach is not needed.…”

“…From the Fourier transform of this region, shown as inserted panels in Figure 7 (mean centering, windowing and zero-padding is applied), the two frequencies are present as two broad local maxima. In order to capture both frequency components of the signal, we propose to include two neighboring frequencies around a number n maxima local maxima, instead of including a portion of the frequencies with the highest amplitude, as suggested by Azarfar et al [16] The effects of including different sets of frequencies are shown in Figure 7, where the sets of selected frequencies are highlighted in the inserted panels. The corresponding corrected spectra are shown in Figure 7, where the frame color matches the color of the corrected spectrum.…”

“…Physical effects are treated as additive contributions, and accounted for by different so-called model spectra. The EMSC model has been developed and combined with meta-modeling of scattering phenomena to handle sophisticated physical phenomena, such as Mie scattering [10,11,13,14] and multiple internal reflection [7,16].…”

“…Fringes are observed in the infrared spectra of many different samples, such as thin tissue sections [7] and paraffin embeddings [15], as well as in measurements with for example microfluidic chambers [16] or diamond anvil cells used in high pressure measurements [17]. Different techniques have been proposed to deal with interference fringes in spectroscopy.…”

“…In EMSC, the amplitude and constant offset are estimated as well. The method, we call it Fringe EMSC, was revised in 2018 by Azafar et al, with the suggestion to make the fringe removal an iterative process, as well as including sine waves of multiple frequencies as model functions in the EMSC [16]. The EMSC based method has the advantage that it can correct many spectra in a fast and robust manner.…”