Neural networks for rapid phase quantification of cultural heritage X‐ray powder diffraction data

• Published in: Journal of applied crystallography Vol. 57; no. 3; pp. 831 - 841
• Main Authors: Poline, Victor, Purushottam Raj Purohit, Ravi Raj Purohit, Bordet, Pierre, Blanc, Nils, Martinetto, Pauline
• Format: Journal Article
• Language: English
• Published: 5 Abbey Square, Chester, Cheshire CH1 2HU, England International Union of Crystallography 01.06.2024; Blackwell Publishing Ltd; International Union of Crystallography / Wiley
• Subjects: Algorithms; Computed tomography; Cultural heritage; Cultural resources; Data acquisition; Data processing; deep learning; Neural networks; Physics; Research Papers; Synchrotron radiation; tomography; X-ray diffraction; computed tomography; X-ray diffraction; deep learning; cultural heritage; tomography; neural networks
• ISSN: 1600-5767; 0021-8898; 1600-5767
• DOI: 10.1107/S1600576724003704

Recent developments in synchrotron radiation facilities have increased the amount of data generated during acquisitions considerably, requiring fast and efficient data processing techniques. Here, the application of dense neural networks (DNNs) to data treatment of X‐ray diffraction computed tomography (XRD‐CT) experiments is presented. Processing involves mapping the phases in a tomographic slice by predicting the phase fraction in each individual pixel. DNNs were trained on sets of calculated XRD patterns generated using a Python algorithm developed in‐house. An initial Rietveld refinement of the tomographic slice sum pattern provides additional information (peak widths and integrated intensities for each phase) to improve the generation of simulated patterns and make them closer to real data. A grid search was used to optimize the network architecture and demonstrated that a single fully connected dense layer was sufficient to accurately determine phase proportions. This DNN was used on the XRD‐CT acquisition of a mock‐up and a historical sample of highly heterogeneous multi‐layered decoration of a late medieval statue, called `applied brocade'. The phase maps predicted by the DNN were in good agreement with other methods, such as non‐negative matrix factorization and serial Rietveld refinements performed with TOPAS, and outperformed them in terms of speed and efficiency. The method was evaluated by regenerating experimental patterns from predictions and using the R‐weighted profile as the agreement factor. This assessment allowed us to confirm the accuracy of the results. A small, fast and efficient neural network for phase fraction prediction of X‐ray diffraction big data is presented. A data‐driven approach allows users to create their own training dataset, making the method fully customizable for each experience.