Baseline correction methods based on penalized least squares are successfully applied to various spectral analyses.

Most spectral data, such as those obtained via infrared, Raman, and mass spectroscopy, have baseline drifts due to fluorescence or other reasons, which have an adverse impact on subsequent analyses. Therefore, several researchers have proposed the use of various baseline-correction methods to address the aforementioned issue. However, most baseline-correction methods require manual adjustment of the parameters to achieve desirable performance. In this study, we propose a baseline-correction method based on a deep-learning model that combines ResNet and UNet. The method uses a deep-learning model trained with simulated spectral data to perform baseline corrections and eliminates the need for manual parameter adjustments. Based on the results of the qualitative and quantitative analyses of the simulated spectral data and actual Raman spectra, the proposed method is easier to apply and has better performance compared to the existing methods. As the proposed method can be applied to Raman spectra and other spectra, it is expected to be widely used.

In this paper, we consider a novel method for identification of Raman spectra recorded on different instruments with different wavelengths.

Raman spectroscopy is widely used for its exceptional identification capabilities in various fields. Traditional methods for target identification using Raman spectroscopy rely on signal correlation with moving windows, requiring data preprocessing that can significantly impact identification performance. In recent years, deep-learning approaches have been proposed to leverage data augmentation techniques, such as baseline and additive noise addition, in order to overcome data scarcity. However, these deep-learning methods are limited to the spectra encountered during training and struggle to handle unseen spectra. To address these limitations, we propose a multi-input hybrid deep-learning model trained with simulated spectral data. By employing simulated spectra, our method tackles the challenges of data scarcity and the handling of unseen spectra encountered in traditional and deep-learning methods. Experimental results demonstrate that our proposed method achieves outstanding identification performance and effectively handles spectra obtained from different Raman spectroscopy systems.

We have proposed a novel deep learning architecture, the Triangular Deep Convolutional Network (TDCN), for Raman spectral baseline correction. TDCN enhances spectral pre-processing, thereby improving the reliability of downstream analyses.