Schematic drawing of the selected pixels, Ai (i = 1, . . , 8), Bi (i = 1, . . , 6), and Ci (i = 1, . . , 3). Pixels Bi and Ci lie between pixels Ai. Scattered radiation of pixel Bi or Ci can be represented as the average of scattered radiation of pixel Ai.
(a) Schematic drawing of the breast phantom (Model 017). (b) The calcification phantom consisting of 42 calcification clusters (Al2O3 crystals) arranged in a 6 × 7 grid. Each cluster consisted of nine calcifications in a 3 × 3 pattern. The clusters were arranged by size into six rows with nominal size ranges of 180–230, 230–280, 280–330, 330–380, 380–430, and 430–480 μm. (c) Imaged object. Calcification phantoms lay on the breast phantom CIRS Model 017.
Schematic of the calibration measurements under narrow-beam geometry.
High energy image of breast phantom, the region inside the green dashed lines is 1000 × 800 pixels, asterisks indicate independent pixels, and dots indicate dependent pixels in algorithmic scatter correction method.
Scatter signal field estimated using the algorithmic method and pinhole-array interpolation method: (a) algorithmic method LE, (b) algorithmic method HE, (c) pinhole-array interpolation method LE, and (d) pinhole-array interpolation method HE.
The differences of SPRs estimated by algorithmic method and pinhole-array method: (a) LE (5 cm thick), (b) HE (5 cm thick), (c) LE (4 cm thick), and (d) HE (4 cm thick).
Scatter signal field estimated using the algorithmic method: (a) sampling interval 50–70 pixels, LE, (b) sampling interval 50–70 pixels, HE, (c) sampling interval 180–210 pixels, LE, and (d) sampling interval 180–210 pixels, HE.
A 80 × 865 pixel corresponding section of the LE and DE calcification images (1643–1722, 68–932) showing calcification size ranges of 180–230 (left), 230–280, 280–330, 330–380, and 380–430 μm (right) for 30% glandularity tissue-equivalent material data. (a) DE/o, (b) DE-pin, (c) DE-al, and (d) DE-al-de.
SPRs measured by the algorithmic method (bold) and pinhole-array interpolation method at the same locations with different glandular ratios in phantoms, low energy.
SPRs measured by the algorithmic method (bold) and pinhole-array interpolation method at the same locations with different glandular ratios in phantoms, high energy.
Background DE calcification signals calculated based on the pinhole-array interpolation scatter correction, the 60 pixel pairs lay on the locations of the center of pinholes.
Background DE calcification signals calculated based on algorithmic scatter correction, the pixel pairs were independent pairs in algorithmic method.
Background DE calcification signals without scatter correction, pixel pairs were the same as in Tables III and IV .
The median, minimum, and maximum background DE calcification signals (μm) in the regions-of-interest with and without scatter correction.
The median, minimum, and maximum background DE calcification signals (μm) in the regions-of-interest when the algorithmic method was used with different sampling intervals.
The median, minimum, and maximum background DE calcification signals (μm) in the regions-of-interest when algorithmic method was used with and without denoising technique. Sampling interval was 60–120 pixels.
A direct comparison of the average calcification CNR of the three calcification sizes (230–280, 280–330, and 330–380 μm) over three glandular ratios of 30%, 45%, and 70% in the LE image and DE calcification images without/with scatter correction (pinhole-array interpolation method and algorithmic method). The CNR values of DE calcification images with algorithmic scatter correction and denoised techniques are also listed.
Article metrics loading...
Full text loading...