Filtered backprojection for modifying the impulse response of circular tomosynthesis

Abstract

A filtering technique has been developed to modify the three-dimensional impulse response of circular motion tomosynthesis to allow the generation of images whose appearance is like those of some other imaging geometries. In particular, this technique can reconstruct images with a blurring function which is more homogeneous for off-focal plane objects than that from circular tomosynthesis. In this paper, we describe the filtering process, and demonstrate the ability to alter the impulse response in circular motion tomosynthesis from a ring to a disk. This filtering may be desirable because the blurred out-of-plane objects appear less structured.

Task-based performance analysis of FBP, SART and ML for digital breast tomosynthesis using signal CNR and Channelised Hotelling Observers

“…Nevertheless, several DBT performance analyses have used CNR as a surrogate metric for lesion detectability (Ren et al, 2005;Stevens et al, 2001;Ludwig et al, 2008;Yoon et al, 2009;Kempston et al, 2006;Diekmann et al, 2007;Zhang et al, 2008;Wu et al, 2003;Wu et al, 2004). In this study, we reiterate that the AUC of a CHO-LG observer is a more appropriate metric, but include CNR results as well to (1) facilitate comparisons with previous studies in the literature, and (2) illustrate the difference in results obtained by the two metrics.…”

“…Several reconstruction methods have been explored for use in DBT, including filtered backprojection (FBP) (Lauritsch and Haerer, 1998;Stevens et al, 2001;Claus et al, 2006;Mertelmeier et al, 2006), algebraic reconstruction (ART) (Kaczmarz, 1937;Andersen et al, 1984;Kak and Slaney, 2001;Kempston et al, 2006;Diekmann et al, 2007;Zhang et al, 2007) and maximum likelihood (ML) techniques (Wu et al, 2003;Wu et al, 2008). Unfortunately, the DBT reconstruction problem is under-defined, meaning that there exists an infinite set of possible solutions that satisfy the projection data, also known as the null space.…”

A comparison between traditional shift-and-add (SAA) and point-by-point back projection (BP) -- relevance to morphology of microcalcifications for isocentric motion in Digital Breast tomosynthesis (DBT)

“…Several tomosynthesis image reconstruction algorithms have been proved effectively in DBT image reconstruction at of yet, including traditional shift-and-add (SAA) [1,2], Niklason and colleagues' image stretching shift-and-add [3], Wu et al's maximum likelihood maximization expectation (MLEM) [4,5], tuned-aperture computed tomography (TACT) [6,7], filtered back projection [2,[16][17][18][19], matrix inversion tomosynthesis [2,8,12], algebraic reconstruction techniques (ART) [14], and Gaussian frequency blending (GFB) of MITS and FBP [9].…”

Statistical iterative reconstruction to improve image quality for digital breast tomosynthesis

“…Since FBP has a linear response, tuning the filter kernel permits a desired trade-off between resolution and noise. As a standard reference method, we applied a sampling density based ramp filter 52 and a Hanning filter to remove the ring effects and high frequency noise. The kernel was applied on each row of projection data in the path direction of the x-ray tube, and then the 3D volume was reconstructed by a pixel-driven backprojection.…”