Vision Transformers for X-ray Diffraction Patterns Analysis
Gespeichert in:
| Titel: | Vision Transformers for X-ray Diffraction Patterns Analysis |
|---|---|
| Autoren: | Simonnet, Titouan, Fall, Mame Diarra, Grangeon, Sylvain, Galerne, Bruno |
| Weitere Verfasser: | COUFFIGNAL, Frédérique |
| Quelle: | ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). :1-5 |
| Verlagsinformationen: | IEEE, 2025. |
| Publikationsjahr: | 2025 |
| Schlagwörter: | Vision Transformers, Deep Learning, [SDU.STU] Sciences of the Universe [physics]/Earth Sciences, XRD patterns |
| Beschreibung: | Understanding materials properties depends largely on the ability to determine its components, and in particular its mineral phases. Powder X-ray diffraction (XRD) is a powerful tool for such purposes. This paper presents a Transformerbased vision model (ViT) for mineral phase identification, and proportion inference to quantify the mineral phases present in a material. Our analysis shows that the tokenization strategy is a critical step for XRD pattern analysis. The results obtained for both tasks are excellent and more robust than those obtained with a CNN. The proposed approach also makes it possible to introduce visualization tools for signal analysis, to better understand how information flows through the model and how data is classified or quantified. |
| Publikationsart: | Article Conference object |
| Dateibeschreibung: | application/pdf |
| DOI: | 10.1109/icassp49660.2025.10887635 |
| Rights: | STM Policy #29 CC BY |
| Dokumentencode: | edsair.doi.dedup.....838794de491adec6a2e7c386f45ccb76 |
| Datenbank: | OpenAIRE |
Nájsť tento článok vo Web of Science | Abstract: | Understanding materials properties depends largely on the ability to determine its components, and in particular its mineral phases. Powder X-ray diffraction (XRD) is a powerful tool for such purposes. This paper presents a Transformerbased vision model (ViT) for mineral phase identification, and proportion inference to quantify the mineral phases present in a material. Our analysis shows that the tokenization strategy is a critical step for XRD pattern analysis. The results obtained for both tasks are excellent and more robust than those obtained with a CNN. The proposed approach also makes it possible to introduce visualization tools for signal analysis, to better understand how information flows through the model and how data is classified or quantified. |
|---|---|
| DOI: | 10.1109/icassp49660.2025.10887635 |