Developing a Peak Extraction and Retention (PEER) Algorithm for Improving the Temporal Resolution of Raman Spectroscopy

Abstract

In spectroscopic analysis, push-to-the-limit sensitivity is one of the important topics, particularly when facing the qualitative and quantitative analyses of the trace target. Normally, the effective recognition and extraction of weak signals are the first key steps, for which there has been considerable effort in developing various denoising algorithms for decades. Nevertheless, the lower the signal-to-noise ratio (SNR), the greater the deviation of the peak height and shape during the denoising process. Therefore, we propose a denoising algorithm along with peak extraction and retention (PEER). First, both the first and second derivatives of the Raman spectrum are used to determine Raman peaks with a high SNR whose peak information is kept away from the denoising process. Second, an optimized window smoothing algorithm is applied to the left part of the Raman spectrum, which is combined with the untreated Raman peaks to obtain the denoised Raman spectrum. The PEER algorithm is demonstrated with much better signal extraction and retention and successfully improves the temporal resolution of Raman imaging of a living cell by at least 1 order of magnitude higher than those by traditional algorithms.

Automatic Identification of Individual Nanoplastics by Raman Spectroscopy Based on Machine Learning

“…Weak Raman signal extraction and denoising are required because the inherent noise of the system interferes with the identification of plastic feature peaks. The PEER algorithm has good signal extraction and retention, as described earlier . To reduce the influence of overlapping peaks and the background on the model construction, the subsequent steps use the first-order derivatives of the Raman spectra.…”

“…The PEER algorithm has good signal extraction and retention, as described earlier. 41 To reduce the influence of overlapping peaks and the background on the model construction, the subsequent steps use the first-order derivatives of the Raman spectra.…”

“…In terms of detection methods, the highly reflective aluminum sheets (95% reflectivity) facilitate automatic image identification of nanoplastics as well as signal collection. For data processing, the peak extraction and retention (PEER) algorithm served to extract meaningful features from weak Raman signals (for technical details see ref ). To identify unknown plastic samples, we constructed a random forest model enabling successful discrimination between five common plastic contaminants, namely, polyethylene (PE), polytetrafluoroethylene (PTFE), polystyrene (PS), poly­(methyl methacrylate) (PMMA), and polyvinyl chloride (PVC), with an average accuracy of 98.8%, an average sensitivity of 98.5%, and an average specificity of 100%.…”

Artificial Intelligence for Surface‐Enhanced Raman Spectroscopy

“…Wavelet decomposition (e.g., wavelet filter) and window moving average (e.g., simple moving average, Savitzky-Golay (SG) filter) are among the two most used denoising algorithms for SERS spectra, but they suffer from intensity loss at shape peaks, insufficient noise reduction, and/or artificial peak production on the heavy-noise region. [272,273] It is possibly because that in these methods, preset parameters function on the whole spectra, but in fact, the noise region and the peak region demand different treatment. On the contrary, ML can automatically differentiate a variety of noise components from the peak after training upon a large dataset (i.e., high and low SNR spectra pairs acquired from the same spot at one sample), thereafter executing smoothing or retention accordingly.…”

Fast Burst-Sparsity Learning-Based Baseline Correction (FBSL-BC) Algorithm for Signals of Analytical Instruments

“…Such a baseline shift could change the essential characteristics of feature peaks (e.g., position, strength, and slope). As a consequence, the collected spectral signals are not strictly adhering to Beer–Lambert Law, which will bring a negative impact on the following qualitative and quantitative analysis. , To avoid the interference of baseline shift, some relevant baseline correction algorithms are required in spectroscopic analysis, either to enhance feature signals with weak intensities or guarantee the accuracy and stability of the calibration model.…”