A Small-Window Moving Average-Based Fully Automated Baseline Estimation Method for Raman Spectra

Abstract

A fully automated and model-free baseline-correction method for vibrational spectra is presented. It iteratively applies a small, but increasing, moving average window in conjunction with peak stripping to estimate spectral baselines. Peak stripping causes the area stripped from the spectrum to initially increase and then diminish as peak stripping proceeds to completion; a subsequent increase is generally indicative of the commencement of baseline stripping. Consequently, this local minimum is used as a stopping criterion. A backup is provided by a second stopping criterion based on the area under a third-order polynomial fitted to the first derivative of the current estimate of the baseline-free spectrum and also indicates whether baseline is being stripped. When the second stopping criterion is triggered instead of the first one, a proportionally scaled simulated Gaussian baseline is added to the current estimate of the baseline-free spectrum to act as an internal standard to facilitate subsequent processing and termination via the first stopping criterion. The method is conceptually simple, easy to implement, and fully automated. Good and consistent results were obtained on simulated and real Raman spectra, making it suitable for the fully automated baseline correction of large numbers of spectra.

Raman spectroscopy of fibroblast cells from a Huntington's disease patient

“…Secondly, the resulting spectra were flattened to remove the effect of cellular autofluorescence using a small-window moving average automated baseline correction (SWiMA) procedure. [34] The Matlab code was kindly provided by the author Schulze. This procedure first applies a window of three points along the wavenumber (x) axis, starting at the left end.…”

Development and Integration of Block Operations for Data Invariant Automation of Digital Preprocessing and Analysis of Biological and Biomedical Raman Spectra

“…40,51,68,70,71,76 Thus, in order to effect data invariant automation, empirically or arbitrarily chosen parameters, stopping criteria, thresholds, and adjustments (i.e., increments, decrements) generally should be avoided. It is important to be When parameter search or estimation is required, it is advisable to use a subset of spectra rather than all the spectra in a potentially significantly larger set.…”

“…Thus either a plateau (no further stripping is occurring) or a local minimum in the sequence of stripped Raman peak areas is indicative that peak stripping has reached completion and that only the baseline remains. 40 Note how initial parameter settings, parameter adjustments, and threshold specifications are based either on extreme values (e.g., largest/smallest, minimum/maximum, etc.) or accepted statistical metrics (v 2 , a = 0.05, etc.…”

Automatic Baseline Recognition for the Correction of Large Sets of Spectra Using Continuous Wavelet Transform and Iterative Fitting

“…Various authors tackled this issue by proposing automated or semi-automated algorithms for baseline correction that reduce human intervention. Effective solutions were devised using iterative polynomial fitting, 13 penalized quantile spline regression, 14 adaptive least squares/ Whittaker smoother, [15][16][17] moving average-peak stripping, [18][19][20] local second derivative, 12 and morphological or geometrical approaches. 21,22 The performance of these methods differ in terms of accuracy, computational speed, amount of human intervention, and types of spectra to which they can be applied; these goals are usually conflicting.…”

“…However, these fully automated methods often require computational times that are too long for certain applications. For example, the algorithm proposed by Schulze et al 19 is reported to correct a single baseline in at least 20 s. For a typical spectral image of 100 3 100 pixels, the corresponding overall correction time is in the order of days. Moreover, the applicability of such methods is limited by the actual difficulty in defining a universal criterion to identify the ''true'' signal that works for any spectrum.…”