Fig 1.
The geometry of digital breast tomosynthesis systems.
(a) step-and-shoot mode and (b) continuous mode. The X-ray source and a flat-panel detector rotate simultaneously along the path within a limited angular range.
Table 1.
Simulation parameters.
Fig 2.
The table-top experimental system.
(a) a Teflon sphere phantom and (b) a BR3D breast phantom.
Table 2.
Experimental parameters.
Table 3.
Inverse filtering approach to measure in-plane MTF.
Fig 3.
Exampled 10 number of LG spatial channel images with au = 11 pixels (left: p = 0, right: p = 9).
Fig 4.
Reconstructed images of spherical signals in (a—c) step-and-shoot mode, and (d—f) continuous mode.
Signal diameters are (a, d) 1 mm, (b, e) 2 mm, and (c, f) 5 mm. The display window is set by [min. max.] mm−1.
Fig 5.
Examples of reconstructed simulated breast phantom without spherical signal.
(a) and (b) are reconstructed images using projection data in step-and-shoot mode, and (c) and (d) are corresponding images using projection data in continuous mode. The display window is set by [min. max.] mm−1 to visualize the background structures more clearly.
Fig 6.
Examples of reconstructed BR3D breast phantom without spherical signal.
(a) is a reconstructed image using projection data in step-and-shoot mode, and (b) is the corresponding image using projection data in continuous mode. Four 100 × 100 images are cropped to show the detailed anatomical background (indicated by the yellow box). The display window is set by [min. max.] mm−1 to visualize the background structures more clearly.
Fig 7.
In-plane MTFs in step-and-shoot mode (indicated by the red color) and continuous mode (indicated by the blue color) using the simulated sphere phantom (indicated by the solid line) and experimental Teflon sphere phantom (indicated by the dashed line).
Fig 8.
Anatomical noise power spectrum from the reconstructed in-plane images using (a) simulated breast phantom and (b) BR3D breast phantom.
Table 4.
β values of the anatomical NPS.
Fig 9.
SNRt values of model observers for reconstructed images of (a) simulated breast phantom and (b) BR3D breast phantom.
The SNRt values of model observers with 95% confidence intervals.
Fig 10.
Examples of reconstructed images with spherical signals from (a) simulated breast phantom and (b) BR3D breast phantom.
In each image set, the left image was obtained in step-and-shoot mode, and the right image was obtained in continuous mode. The display window is set by [min. max.] mm−1 to visualize the background structures more clearly.