Offre de thèse : Development of machine learning methods for data analysis of X-ray diffraction mapping applied to cultural heritage (Grenoble/Antwerp)


X-ray diffraction mapping (combined with X-ray fluorescence mapping) is increasingly used to study cultural heritage (CH) artefacts. Synchrotron-based micro-beams are well suited for the analysis of small samples while portable and Inverse Compton Scattering (ICS) sources -based instruments have been developed for the analysis of entire CH objects such as paintings. In both cases, the obtained crystal phase maps can be exploited to obtain information about the artistic techniques as well as the conservation state of the CH artefacts. The development of new X-ray sources and new detectors permit a dramatic improvement of acquisition speed. The main bottleneck is now data analysis which can be either fast but qualitative (based on XRD peak intensities) or quantitative (Rietveld refinement) but slow. Machine learning is a promising approach for addressing this challenge.

Plus d’informations :
[Télécharger l'annonce - PDF 150 Ko] et [Website ENGAGE project]