Deep Learning for Reconstructing Low-Quality FTIR and Raman Spectra─A Case Study in Microplastic Analyses

Abstract

Herein we report on a deep-learning method for the removal of instrumental noise and unwanted spectral artifacts in Fourier transform infrared (FTIR) or Raman spectra, especially in automated applications in which a large number of spectra have to be acquired within limited time. Automated batch workflows allowing only a few seconds per measurement, without the possibility of manually optimizing measurement parameters, often result in challenging and heterogeneous datasets. A prominent example of this problem is the automated spectroscopic measurement of particles in environmental samples regarding their content of microplastic (MP) particles. Effective spectral identification is hampered by low signal-to-noise ratios and baseline artifacts as, again, spectral post-processing and analysis must be performed in automated measurements, without adjusting specific parameters for each spectrum. We demonstrate the application of a simple autoencoding neural net for reconstruction of complex spectral distortions, such as high levels of noise, baseline bending, interferences, or distorted bands. Once trained on appropriate data, the network is able to remove all unwanted artifacts in a single pass without the need for tuning spectra-specific parameters and with high computational efficiency. Thus, it offers great potential for monitoring applications with a large number of spectra and limited analysis time with availability of representative data from already completed experiments.

Cascaded Deep Convolutional Neural Networks as Improved Methods of Preprocessing Raman Spectroscopy Data

“…Additionally, they have been used successfully for Raman and FTIR spectroscopy applications to preprocess the raw data. 37 Ideally, the reduction of the input's dimensionality does not leave enough room for noise such that the trained network's output will effectively denoised. In the proposed network, as the output is not identical to its input, it is not considered a conventional autoencoder.…”

Accurate Tumor Subtype Detection with Raman Spectroscopy via Variational Autoencoder and Machine Learning

“…Some studies have been conducted using autoencoders (e.g., sparse autoencoder) to process spectra for improved analysis in fields such as tumor types, microplastics, and drug resistance. 27 − 29 However, the application of autoencoders for classifying tumor subtypes is still lacking. VAE encodes Raman spectra as two-dimensional (2D) information, which fits a Gaussian distribution, and the fewer data dimensions make it insufficient to store noise information.…”

“…The variational autoencoder (VAE) was employed to reduce feature redundancy and to accomplish noise reduction. Some studies have been conducted using autoencoders (e.g., sparse autoencoder) to process spectra for improved analysis in fields such as tumor types, microplastics, and drug resistance. − However, the application of autoencoders for classifying tumor subtypes is still lacking. VAE encodes Raman spectra as two-dimensional (2D) information, which fits a Gaussian distribution, and the fewer data dimensions make it insufficient to store noise information. − Nine classification models were trained to classify the VAE-encoded Raman spectra of three subtypes of non-small cell lung cancer (NSCLC) cell lines (A549, H1299, and H460) and two subtypes of kidney cancer tissues (clear cell renal cell carcinomas (CCRCC) and papillary renal cell carcinoma (PRCC)).…”

Advanced microplastic monitoring using Raman spectroscopy with a combination of nanostructure-based substrates

“…The detection of MNPs require more consideration and a focus on novel detection methods, such as plasmonic resonance-based Raman spectroscopy sensors. Interestingly, spectroscopic methods that have been well investigated and shown to provide highly sensitive trace detection of MNPs include Raman scattering microscopy methods, such as Raman imaging, normal Raman spectrum [ 73 , 74 ], deep learning for reconstructing low-quality FTIR and Raman spectra [ 75 ], or the combination of Raman imaging and matrix-assisted laser desorption/ionization mass spectrometry [ 76 ], thermal gravimetric analysis, FTIR spectroscopy, or gas chromatography mass spectrometry [ 77 ]. Recently, metal plasmonic nanostructures have emerged as potential materials utilized in various fields such as energy application [ 78 , 79 ], catalytic reduction [ 80 ], and efficient dye removal [ 81 ].…”