Recent progress and applications of Raman spectrum denoising algorithms in chemical and biological analyses: A review

Raman spectroscopy is a powerful technique widely used in analytical chemistry. However, spectral noise emerging during detection introduces potential to compromise the signal-to-noise ratio, undermining the accuracy and reliability of sample analyses. This limitation has driven the development of Raman spectrum denoising algorithms to encourage the application of Raman spectroscopy in more complicated domains of analytical chemistry, including unknown compound identification, trace detection, single-particle sensing, ultrafast imaging, and in-depth in vivo detections. In this review, we outline the essential concepts of Raman spectroscopy, the origins of spectral noise and noise reduction through various strategies. Furthermore, we present a comprehensive summary of Raman spectral denoising algorithms and their progressions in three categories of moving window smoothing, power spectrum estimation, and deep learning algorithms. Finally, challenges and future directions in denoising algorithms are discussed. This review severs as a valuable resource to shed light on algorithmic solutions to enhance Raman spectroscopy analysis.