TY - JOUR

T1 - Spectrum is a picture: Feasibility study of two-dimensional convolutional neural networks in spectral processing

AU - Deev, V.

AU - Panchuk, V.

AU - Boichenko, E.

AU - Kirsanov, D.

N1 - Export Date: 5 October 2024 CODEN: MICJA Адрес для корреспонденции: Kirsanov, D.; Institute of Chemistry, Peterhof, Universitetsky Prospect, 26, Russian Federation; эл. почта: d.kirsanov@gmail.com Сведения о финансировании: 23-73-01139 Сведения о финансировании: Ministry of Education and Science of the Russian Federation, Minobrnauka, FFZM-2022-0008, 2 542,089 Сведения о финансировании: Ministry of Education and Science of the Russian Federation, Minobrnauka Текст о финансировании 1: The authors are grateful to A. Panchenko and M. Tyndyk (National Medical Research Center of Oncology N.N. Petrov, St. Petersburg, Russia) for the provided biological materials. E. Boichenko acknowledges financial support of the NIR studies from RSF project #23-73-01139. V. Panchuk acknowledges financial support of M\u00F6ssbauer studies from Ministry of Education and Science of the Russian Federation, Project \u2116 FFZM-2022-0008 (Subject \u2116 2 542,089).

PY - 2024/10/1

Y1 - 2024/10/1

N2 - In chemometrics, a spectrum is usually represented in a form of numeric vector for further processing. The same approach has been used also for spectral data treatment by convolutional neural networks (CNN), initially purposed, however, for image processing. Analyzing spectral data as a two-dimensional picture rather than a one-dimensional vector potentially can improve the accuracy of regression and classification models. The purpose of this work was to test this assumption. As a first case study we have chosen one of the most difficult types of spectra to interpret – the Mössbauer spectral data. We explored the potential of 2D-CNN to predict Mössbauer spectral parameters numerically using image (.bmp) files with spectra. As a second case study, we address another challenging task – classification of biological tissues using their near-infrared spectra. In both cases, we compared the performance of CNN for two types of input data: spectra as pictures and as numeric vectors. It was found that in case of Mossbauer spectra, the use of 2D-CNN provides for better prediction of spectral parameters, while for NIR spectra of biological tissues the use of the traditional PLS-DA method turned out to be better for classification compared to the CNN approach. © 2024 Elsevier B.V.

AB - In chemometrics, a spectrum is usually represented in a form of numeric vector for further processing. The same approach has been used also for spectral data treatment by convolutional neural networks (CNN), initially purposed, however, for image processing. Analyzing spectral data as a two-dimensional picture rather than a one-dimensional vector potentially can improve the accuracy of regression and classification models. The purpose of this work was to test this assumption. As a first case study we have chosen one of the most difficult types of spectra to interpret – the Mössbauer spectral data. We explored the potential of 2D-CNN to predict Mössbauer spectral parameters numerically using image (.bmp) files with spectra. As a second case study, we address another challenging task – classification of biological tissues using their near-infrared spectra. In both cases, we compared the performance of CNN for two types of input data: spectra as pictures and as numeric vectors. It was found that in case of Mossbauer spectra, the use of 2D-CNN provides for better prediction of spectral parameters, while for NIR spectra of biological tissues the use of the traditional PLS-DA method turned out to be better for classification compared to the CNN approach. © 2024 Elsevier B.V.

KW - Classification

KW - CNN

KW - Deep learning

KW - Mössbauer spectroscopy

KW - Regression

UR - https://www.mendeley.com/catalogue/cf388c71-7532-329b-996f-b9ddabd1e0b5/

U2 - 10.1016/j.microc.2024.111329

DO - 10.1016/j.microc.2024.111329

M3 - статья

VL - 205

JO - Microchemical Journal

JF - Microchemical Journal

SN - 0026-265X

M1 - 111329

ER -