Machine learning tools for the classification of visible reflectance spectra of inorganic pigments historically used in wall paintings

Jordi Cruz; Roberto Sáez Hernández

doi:10.37558/gec.v28i1.1444

Autores

Jordi Cruz EUSS School of Engineering, Barcelona https://orcid.org/0000-0001-8191-8689
Roberto Sáez Hernández Universitat de València https://orcid.org/0000-0001-9010-9445

DOI:

https://doi.org/10.37558/gec.v28i1.1444

Palavras-chave:

Artificial Neural Networks, Classification, Chemometrics, Identification, Partial Least Squares Discriminant Analysis, Pigments, Support Vector Machines, Visible reflectance spectroscopy

Resumo

Three different classification methods applied to the field of Cultural Heritage are compared. Focusing on the non-invasive and in situ identification of historical inorganic pigments using visible reflectance spectroscopy, linear and non-linear classifiers are trained and evaluated (Partial Least Squares – Discriminant Analysis (PLS-DA), Support Vector Machine (SVM-DA), and Artificial Neural Networks (ANN-DA)). Three hues (yellow, blue and green) are considered with three pigments per hue: lead-tin yellow, yellow ochre, orpiment (yellow); celadonite, green malachite, Verona green (green); and Egyptian blue, azurite, and lapis lazuli (blue). All three classification methods yielded high prediction capabilities (above 94.2 % in the external validation sets), with high sensitivity and specificity (above 88 %). In cases where spectral features are more similar, PLS-DA performed slightly worse. However, SVM-DA and ANN-DA models were comparable in all instances, with global accuracies above 98.8 %, sensitivity above 88 %, and specificity above 98 %, respectively.

Downloads

Não há dados estatísticos.

Biografias Autor

Jordi Cruz, EUSS School of Engineering, Barcelona

Jordi Cruz is a University Associate Professor and lecturer at the EUSS School of Engineering in Barcelona since 2011. He holds an h-index of 11 and has authored over 30 scientific publications. He has supervised two PhD theses and one master’s student. His research focuses on multivariate analysis applied to rapid analytical techniques such as NIR, Raman, FTIR, and NIR chemical imaging spectroscopy. These methods are used to develop fast, reliable analytical tools across various industrial sectors, including pharmaceuticals, phytotherapy, petrochemicals, cultural heritage, and forensics.
Dr. Cruz conducts most of his research at the Agri-Food Technology Research Institute (IRTA) in Monells (Girona) and Forensics and Cultural Heritage at SmartChem Lab at the University of Valencia. He also maintains active collaborations with several international institutions: the Department of Pharmaceutical Chemistry at Mahidol University (Bangkok, Thailand), the Institute of Chemical Engineering at the Bulgarian Academy of Sciences (Sofia, Bulgaria), the University of the Basque Country, Teagasc in Ireland, and the Department of Forensic Chemistry at Florida International University (USA).

Roberto Sáez Hernández, Universitat de València

Roberto Sáez-Hernández is an Assistant Professor at the Analytical Chemistry department of the University of Valencia. His research focuses on the analytical treatment and interpretation of spectroscopic and imaging signals through chemometric exploratory and supervised methods. Particularly, he has focused his research on the fields of food science – delving into the investigation of food contamination and food fraud –, Cultural Heritage – investigating construction materials of relevant historical sites or ancient gold coinage –, and more recently Forensics. In this regard, he has carried out research stays in national and international universities and research facilities, like the Institute of Food Science and Technology (IATA – CSIC), and the Universities of Bologna, Autonomous of Barcelona, and Copenhagen

Referências

ACETO, M., APOLLONIA, L., & LAMBERTI, A. (2014). “Characterisation of colourants on illuminated manuscripts by portable fibre optic UV–visible–NIR reflectance spectrophotometry”, Analytical Methods, 6(5): 1488–1500. https://doi.org/10.1039/C3AY41904E DOI: https://doi.org/10.1039/c3ay41904e

BALLABIO, D., & CONSONNI, V. (2013). “Classification tools in chemistry. Part 1: linear models. PLS-DA”, Analytical Methods, 5(16): 3790. https://doi.org/10.1039/c3ay40582f DOI: https://doi.org/10.1039/c3ay40582f

BARBU, O.H. (2021). “pXRF and FTIR Spectrometry Applied to the Study of Azurite and Smalt in Romanian Medieval Wall Painting”, In: Aoki, S., et al. “Conservation and Painting Techniques of Wall Paintings on the Ancient Silk Road.” Cultural Heritage Science. Springer, Singapore. https://doi.org/10.1007/978-981-33-4161-6_6 DOI: https://doi.org/10.1007/978-981-33-4161-6_6

BIANCHETTI, P., TALARICO, F., VIGLIANO, M. G., & ALI, M. F. (2000). “Production and characterization of Egyptian blue and Egyptian green frit”, Journal of Cultural Heritage, 1(2): 179–188. https://doi.org/10.1016/s1296-2074(00)00165-5 DOI: https://doi.org/10.1016/S1296-2074(00)00165-5

CHEILAKOU, E., KARTSONAKI, M., KOUI, M., & CALLET, P. (2009). “A nondestructive study of the identification of pigments on monuments by colorimetry”, International Journal of Microstructure and Materials Properties, 4(1): 112. https://doi.org/10.1504/ijmmp.2009.028437 DOI: https://doi.org/10.1504/IJMMP.2009.028437

CSÉFALVAYOVÁ, L., STRLIČ, M., KARJALAINEN, H. (2011). “Quantitative NIR Chemical Imaging in Heritage Science”, Analytical Chemistry, 83(13): 5101 – 5106. https://doi.org/10.1021/ac200986p DOI: https://doi.org/10.1021/ac200986p

DASZYKOWSKI, M., WALCZAK, B., & MASSART, D. (2002). “Representative subset selection”, Analytica Chimica Acta, 468(1): 91–103. https://doi.org/10.1016/s0003-2670(02)00651-7 DOI: https://doi.org/10.1016/S0003-2670(02)00651-7

DELANEY, J., RICCIARDI, P., & ACETO, M. (2014). “Use of imaging spectroscopy, fiber optic reflectance spectroscopy, and X-ray fluorescence to map and identify pigments in illuminated manuscripts”, Heritage Science, 2(6). https://doi.org/10.1179/2047058412Y.0000000078 DOI: https://doi.org/10.1179/2047058412Y.0000000078

FAN, C., ZHANG, P., WANG, S., & HU, B. (2018). “A study on classification of mineral pigments based on spectral angle mapper and decision tree”, Proceedings of SPIE, 10806, Article 10806. https://doi.org/10.1117/12.2503088 DOI: https://doi.org/10.1117/12.2503088

FIORETTI, G., CLAUSI, M., ERAMO, G., LONGO, E., MONNO, A., PINTO, D., & TEMPESTA, G. (2023). “A Non-Invasive and sustainable characterization of pigments in wall paintings: a library of Apulian colors”, Heritage, 6(2): 1567–1593. https://doi.org/10.3390/heritage6020084 DOI: https://doi.org/10.3390/heritage6020084

GARCÍA-BUCIO, M. A., CASANOVA-GONZÁLEZ, E., MITRANI, A., RUVALCABA-SIL, J. L., MAYNEZ-ROJAS, M. Á., & RANGEL-CHÁVEZ, I. (2022). “Non-destructive and non-invasive methodology for the in situ identification of Mexican yellow lake pigments”, Microchemical Journal, 183, 107948. https://doi.org/10.1016/j.microc.2022.107948 DOI: https://doi.org/10.1016/j.microc.2022.107948

HRADIL, D., GRYGAR, T., HRADILOVÁ, J., BEZDIČKA, P., GRŰNWALDOVÁ, V., FOGAŠ, I., & MILIANI, C. (2007). “Microanalytical identification of Pb-Sb-Sn yellow pigment in historical European paintings and its differentiation from lead tin and Naples yellows”, Journal of Cultural Heritage, 8(4): 377–386. https://doi.org/10.1016/j.culher.2007.07.001 DOI: https://doi.org/10.1016/j.culher.2007.07.001

KASTENHOLZ, H.V., TOPPER, M.I., WARREN, W.S., FISCHER, M.C., & GRASS, D. (2024). “Noninvasive identification of carbon-based black pigments in cultural heritage using pump–probe microscopy and support vector machines”, Proceedings of the National Academy of Sciences, 121(50), e2407433121. https://doi.org/10.1126/sciadv.adp0005 DOI: https://doi.org/10.1126/sciadv.adp0005

LIU, M., WANG, Z., & LIU, X. (2024). “Spectroscopic Techniques for Identifying Pigments in Polychrome Cultural Relics”, Coatings, 15(1): 20. https://doi.org/10.3390/coatings15010020 DOI: https://doi.org/10.3390/coatings15010020

MANDAL, D., PEDERSEN, M., GEORGE, S., DEBORAH, H., & BOUST, C. (2023). “An Experiment-based Comparative Analysis of Pigment Classification Algorithms using Hyperspectral Imaging”, Journal of Imaging Science and Technology, 67(3), 030403. https://doi.org/10.2352/j.imagingsci.technol.2023.67.3.030403 DOI: https://doi.org/10.2352/J.ImagingSci.Technol.2023.67.3.030403

MAO, J., & JAIN, A. (2002). “Discriminant analysis neural networks”, IEEE International Conference on Neural Networks, San Francisco: IEEE, 1993, 300-305 vol.1. https://doi.org/10.1109/icnn.1993.298573 DOI: https://doi.org/10.1109/ICNN.1993.298573

MELO, M.J., NABAIS, P., VIEIRA, M., ARAÚJO, R., OTERO, V., LOPES, J., & MARTÍN, L. (2022). “Between past and future: Advanced studies of ancient colours to safeguard cultural heritage and new sustainable applications”, Dyes and Pigments, 205, 110563. https://doi.org/10.1016/j.dyepig.2022.110815 DOI: https://doi.org/10.1016/j.dyepig.2022.110815

MONTAGNER, C., SANCHES, D., PEDROSO, J., MELO, M. J., & VILARIGUES, M. (2012). “Ochres and earths: Matrix and chromophores characterization of 19th and 20th century artist materials”, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 103: 409–416. https://doi.org/10.1016/j.saa.2012.10.064 DOI: https://doi.org/10.1016/j.saa.2012.10.064

MORETTO, L. M., ORSEGA, E. F., & MAZZOCCHIN, G. A. (2011). “Spectroscopic methods for the analysis of celadonite and glauconite in Roman green wall paintings”, Journal of Cultural Heritage, 12(4): 384–391. https://doi.org/10.1016/j.culher.2011.04.003 DOI: https://doi.org/10.1016/j.culher.2011.04.003

MUÑOZ, L.P. (2023). “Non-invasive, non-destructive and portable identification of Indian yellow: Moving from highly purified standards to real watercolour, oil and historical samples”, Journal of Cultural Heritage, 62: 202–213. https://doi.org/10.1016/j.culher.2023.06.022 DOI: https://doi.org/10.1016/j.culher.2023.06.022

MUEHLETHALER, C., MASSONNET, G., ESSEIVA, P., & LAMBERT, D. (2016). “Combining spectroscopic data in the forensic analysis of paint: Application of a multiblock technique as chemometric tool”, Forensic Science International, 266: 338–345. https://doi.org/10.1016/j.forsciint.2016.03.049 DOI: https://doi.org/10.1016/j.forsciint.2016.03.049

PIOVESAN, R., MAZZOLI, C., MARITAN, L., CORNALE, P. (2012). “Fresco and lime-paint: an experimental study and objective criteria for distingushing between these painting techniques”, Archaeometry, 54(4): 723 – 736 DOI: https://doi.org/10.1111/j.1475-4754.2011.00647.x

POPELKA-FILCOFF, R.S. (2020). “Pigment Analysis in Archaeology”, En Encyclopedia of Global Archaeology, 8615 – 8619, Springer. https://doi.org/10.1007/978-3-030-30018-0_2281 DOI: https://doi.org/10.1007/978-3-030-30018-0_2281

POUYET, E., MITEVA, T., ROHANI, N., & VIGUERIE, L. (2021). “Artificial Intelligence for Pigment Classification Task in the Short-Wave Infrared Range”, Sensors, 21, Article 6150. https://doi.org/10.3390/s21186150 DOI: https://doi.org/10.3390/s21186150

RADPOUR, R., GATES, G.A., KAKOULLI, I., DELANEY, J.K. (2022). “Identification and mapping of ancient pigments in a Roman Egyptian funerary portrait by application of reflectance and luminescence imaging spectroscopy”, Heritage Science, 10, 8. https://doi.org/10.1186/s40494-021-00639-5 DOI: https://doi.org/10.1186/s40494-021-00639-5

RICCIARDI, P., DELANEY, J.K., & ACETO, M. (2013). “‘It’s not easy being green’: a spectroscopic study of verdigris and copper-containing green pigments in illuminated manuscripts”, Analytical Methods, 5(13): 3278–3288. https://doi.org/10.1039/C3AY40530C DOI: https://doi.org/10.1039/c3ay40530c

RICCIARDI, P., CORDERO, R., & LAMBERTI, A. (2012). “Near Infrared Reflectance Imaging Spectroscopy to Map Paint Binders in Early Renaissance Paintings”, Angewandte Chemie International Edition, 51(32): 7608–7611. https://doi.org/10.1002/anie.201200840 DOI: https://doi.org/10.1002/anie.201200840

ROMANI, M., CAPOBIANCO, G., PRONTI, L., COLAO, F., SECCARONI, C., PUIU, A., FELICI, A.C., VERONA-RINATI, G., CESTELLI-GUIDI, M., TOGNACCI, A., VENDITTELLI, M., MANGANO, M., ACCONCI, A., BONIFAZI, G., SERRANTI, S., MARINELLI, M., & FANTONI, R. (2020). “Analytical chemistry approach in cultural heritage: the case of Vincenzo Pasqualoni’s wall paintings in S. Nicola in Carcere (Rome)”, Microchemical Journal, 156, 104831. https://doi.org/10.1016/j.microc.2020.104920 DOI: https://doi.org/10.1016/j.microc.2020.104920

SÁEZ-HERNÁNDEZ, R., ANTELA, K. U., GALLELLO, G., et al. (2022). “Identification of ancient pigments using non-destructive spectroscopic techniques: a case study on Roman frescoes”, Journal of Cultural Heritage, 58: 156–166. https://doi.org/10.1016/j.culher.2022.10.003 DOI: https://doi.org/10.1016/j.culher.2022.10.003

SÁEZ-HERNÁNDEZ, R., CERVERA, M. L., MORALES-RUBIO, Á., LUQUE, M. J., PÉREZ-TORRALBA, I., GALLELLO, G., ANTELA, K. U., & MAURI-AUCEJO, A. R. (2024). “Digital image-based method to identify historical pigments in wall paintings”, Dyes and Pigments, 222, 111912. https://doi.org/10.1016/j.dyepig.2023.111912 DOI: https://doi.org/10.1016/j.dyepig.2023.111912

VITRUVIO POLION, M.L., “Los diez libros de Arquitectura”.

WU, T, LI, G., YANG, Z., ZHANG, H., LEI, Y., WANG, N., ZHANG, L (2016). “Shortwave Infrared Imaging Spectroscopy for Analysis of Ancient Paintings”, Applied Spectroscopy, 71(5): 977 – 987. https://doi.org/10.1177/0003702816660724 DOI: https://doi.org/10.1177/0003702816660724

YOGURTCU, B., CEBI, N., KOÇER, A.T., & ERARSLAN, A. (2024). “A Review of Non-Destructive Raman Spectroscopy and Chemometric Techniques in the Analysis of Cultural Heritage”, Molecules, 29(22): 5324. https://doi.org/10.3390/molecules29225324 DOI: https://doi.org/10.3390/molecules29225324