Variational Mode Decomposition for Raman Spectral Denoising

An automatic method for accurate signal-to-noise ratio estimation and baseline correction of Raman spectra of environmental microplastics

In this study, we introduced a k-iterative double sliding-window (DSW^k) method for the estimation of spectral noise, signal-to-noise ratio (SNR), and baseline correction. The performance was evaluated using simulated spectra and compared against other commonly employed methods. Convergent evaluation determined that a k value of 20 strikes an optimal balance between convergence and computational intensity. The DSW^k method demonstrated outstanding performance across different spectral types (flat baseline, baseline with elevation, baseline with fluctuation, baseline with elevation and fluctuation) coupled with SNR values from 10 to 1000, achieving results that ranged from 1.01 to 1.08 times of the reference value in estimating spectral noise. It also showed that the estimated SNR values are 0.89 to 0.93 times of the reference value, demonstrating a 74.5 % - 131.7 % improvement over the conventional method in spectra with elevated and/or fluctuating baselines. Additionally, the DSW^k method proved effective in correcting baselines and identifying polymers in environmental samples of polyethylene (PE), polypropylene (PP), and polystyrene (PS), despite the limitation of reducing the peak height in spectra with low SNR. This method offers the potential to enhance the automatic and accurate evaluation of spectral quality and could assist in the development of guidelines for more rapid parameter adjustments in Raman measurements.

Improving signal-to-noise ratio of Raman measurements based on ensemble learning approach

Raman spectroscopy is an extensively explored vibrational spectroscopic technique to analyze the biochemical composition and molecular structure of samples, which is often assumed to be non-destructive when carefully using proper laser power and exposure time. However, the inherently weak Raman signal and concurrent fluorescence interference often lead to Raman measurements with a low signal-to-noise ratio (SNR), especially for biological samples. Great efforts have been made to develop experimental approaches and/or numerical algorithms to improve the SNR. In this study, we proposed an ensemble learning approach to recover and denoise Raman measurements with a low SNR. The proposed ensemble learning approach was evaluated on 986 pairs of Raman measurements, each pair of which consists of a low SNR Raman spectrum and a high SNR reference Raman spectrum from the exact same fungal sample but uses 200 times the integration time. Compared with conventional methods, the Raman measurements recovered by the proposed ensemble learning approach are more identical to high SNR reference Raman measurements, with an average RMSE and MAE of only 1.337 × 10-2 and 1.066 × 10-2, respectively; thus, the proposed ensemble learning approach is expected to be a powerful tool for numerically improving the SNR of Raman measurements and further benefits rapid Raman acquisition from biological samples.

Enhancing spatial resolution in Fourier transform infrared spectral image via machine learning algorithms

Owing to the intrinsic signal noise in the characterization of chemical structures through Fourier transform infrared (FT-IR) spectroscopy, the determination of the signal-to-noise ratio (SNR) depends on the level of the concentration of the chemical structures. In situations characterized by limited concentrations of chemical structures, the traditional approach involves mitigating the resulting low SNR by superimposing repetitive measurements. In this study, we achieved comparable high-quality results to data scanned 64 times and superimposed by employing machine learning algorithms such as the principal component analysis and non-negative matrix factorization, which perform the dimensionality reduction, on FT-IR spectral image data that was only scanned once. Furthermore, the spatial resolution of the mapping images correlated to each chemical structure was enhanced by applying both the machine learning algorithms and the Gaussian fitting simultaneously. Significantly, our investigation demonstrated that the spatial resolution of the mapping images acquired through relative intensity is further improved by employing dimensionality reduction techniques. Collectively, our findings imply that by optimizing research data through noise reduction enhancing spatial resolution using the machine learning algorithms, research processes can be more efficient, for instance by reducing redundant physical measurements.