Spectral multi-material retrieval using edge illumination single mask X-ray phase contrast imaging

This study presents a two-dimensional computational simulation in Geant4 to quantitatively implement an X-ray spectral multi-material decomposition approach with the edge illumination single mask method. The aim is to simultaneously retrieve the projected thickness maps of different materials in two multi-material samples and directly estimate their phase shift information. Images in multiple energy bins (i.e., energy-bin images) are generated from a simulated monochromatic X-ray source and a weighted polychromatic spectrum. By analyzing the contrast-to-noise ratio (CNR) in each energy-bin image, the optimal energy ranges for inclusion in the spectral approach are identified. The results demonstrate that the approach successfully retrieves the projected thickness maps, showing trends consistent with the theoretical thickness maps. Furthermore, the projected thickness maps for each material are retrieved independently, effectively isolating each material's contribution even in cases of overlap. By using the obtained projected thickness maps, the phase shift map can be derived. Compared to the phase shift map estimated with the Wiener approach, the proposed approach provides a more accurate phase distribution and remains artifact-free across the entire image.