| Egyptian Journal of Archeological and Restoration Studies


EJARS (Established 2011)

Volume 13 ,issue 2 | Summer and Autumn 2023 | Pages : 265-278

DOCUMENTING COLORS BY SPECTRAL FINGERPRINT IN OIL PAINTINGS USING ARTIFICIAL INTELLIGENCE

Document Type: Original Article

DOI: 10.21608/ejars.2023.330909

pages: 265-278

Authors:
Abo Taleb, Th.
Conservation dept., Faculty of Archaeology, Aswan Univ., Aswan, Egypt

Abstract:
Technical documentation helps date the painting and determine the restoration method. Artistic documentation includes the knowledge of the materials used in drawing and the artistic style. The artistic style was analyzed using the spectral fingerprint of color and texture based on the data of the spectral reflectance of each shade on the surface of the painting, especially a distinct area in which the texture of the brush appeared. This was done using Python and artificial intelligence-related computer programs. The result showed that the number of colors was nine with wavelengths of (613.48, 611.37, 594.01, 587.1, 599.42, 497.42, 603.03, 481.94, and 566.33) nm. The texture of the brush was hard and with parallel lines. Sim-ilarly, analyses and examinations were conducted on the painting to determine its components. They helped in the identify the artist and the period. Additionally, in the event of a change in color during the preservation intervention was made, and the environment provided was suitable. The study was conducted on a private oil painting representing impressionism in Europe. It was found that the painting suffered from many aspects of deterioration. XRD-FTIR-Raman (SEM-EDX) analysis showed the use of iron oxide in skin color, zinc oxide additives to obtain color gradations, and chromium green mixed with a small percentage of ground green to obtain gradations in the hilly area of the painting. FTIR and Raman analysis showed that the medium used was linseed oil. Furthermore, the microscopy and stereomicroscope examination of the cross-section showed the cross-sectional structure, the number of imaging layers, and several aspects of damage. Thus, an archival record was produced for future monitoring to prioritize treatment aspects, develop a current pre-treatment record, determine the treatment needed, and provide information on the treatment materials and methods used.

Keywords:
Artificial intelligence ID COLOER Archiving Digitization Computer vision Python

Statistics :
Article views: 432
PDF Downloads: 98