Ultra-sparse View X-ray Computed Tomography for 4-D Imaging

X-ray Computed Tomography (CT) is a non-invasive, non-destructive approach to imaging materials, material systems and engineered components in two- and three- dimensions. Acquisition of 3D images requires the collection of hundreds or thousands of through-thickness X-ray radiographic images from different angles. Such 3D data acquisition strategies commonly involve sub-optimal temporal sampling for in situ and operando studies (4D imaging). Herein, we introduce a sparse-imaging approach, Tomo-NeRF, which is capable of reconstructing high-fidelity 3D images from <10 two-dimensional radiographic images. Experimental 2D and 3D X-ray images were used to test the reconstruction capability in two-view, four-view, and six-view scenarios. Tomo-NeRF is capable of reconstructing 3D images with a structural similarity of 0.9971-0.9975 and voxel-wide accuracy of 81.83–89.59% from 2-D experimentally obtained images. The reconstruction accuracy for the experimentally obtained images is less than the synthetic structures which demonstrated a similarity of 0.9973-0.9984 and voxel-wise accuracy of 84.31-95.77%