Accelerated iterative image reconstruction methods based on block-circulant system matrix derived from a cylindrical image representation

Leroux, J.-D.; Selivanov, V.; Fontaine, R.

doi:10.1109/nssmic.2007.4436714

Cited by 9 publications

(7 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A disadvantage of using a simple azimuthal division of the FOV is the resulting radial degradation of the voxels size. Some excellent alternatives to this problem were proposed in PET, [17][18][19][20] but cannot be straightforwardly applied to CT FOVs, as they do not respect the required number of azimuthal voxels needed for the CT symmetries to exist. An easy workaround is a meticulous choice of the polar grid discretization [see Fig.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Fully 3D iterative CT reconstruction using polar coordinates

2013

Self Cite

View full text Add to dashboard Cite

show abstract

Section: Discussionmentioning

confidence: 99%

“…The use of polar coordinates was already investigated in PET (Refs. [14][15][16][17][18][19][20] and SPECT (Refs. [21][22][23] reconstruction, but has received less attention in CT given the little interest for iterative reconstruction up until recently.…”

Section: Introductionmentioning

confidence: 99%

Fully 3D iterative CT reconstruction using polar coordinates

2013

Self Cite

View full text Add to dashboard Cite

show abstract

“…Examples of these matrices can be found in applications involving discretization aware of cylindrical or cyclical symmetries in the domain [4,9,13,20]. Particularly in the context of CT scanners, mathematical descriptions based on polar coordinates take advantage of the rotational symmetries in the tomographic system, because it is easy to find an ordering of the unknowns so that projections share the same pattern of weights in the probability matrix, although shifted by a fixed number of columns [19,16].…”

Section: Introductionmentioning

confidence: 99%

A Fast Sparse Block Circulant Matrix Vector Product

Romero

Tomás

Soriano

et al. 2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. In the context of computed tomography (CT), iterative image reconstruction techniques are gaining attention because high-quality images are becoming computationally feasible. They involve the solution of large systems of equations, whose cost is dominated by the sparse matrix vector product (SpMV). Our work considers the case of the sparse matrices being block circulant, which arises when taking advantage of the rotational symmetry in the tomographic system. Besides the straightforward storage saving, we exploit the circulant structure to rewrite the poor-performance SpMVs into a high-performance product between sparse and dense matrices. This paper describes the implementations developed for multi-core CPUs and GPUs, and presents experimental results with typical CT matrices. The presented approach is up to ten times faster than without exploiting the circulant structure.

show abstract

“…Monte Carlo system modeling ͑MCSM͒ methods do not have these limitations. These methods have the potential to model both system-dependent factors [15][16][17][18][19][20][21] and object-dependent factors. 22 However, the computational load to calculate the SM and the memory capacity to store the SM have limited the use of these methods for state-of-the-art whole-body PET scanners.…”

Section: Introductionmentioning

confidence: 99%

“…7, 23 To address these problems, symmetries have been exploited to reduce the redundancy in the SM with voxels as the object model. 17,24 Alternative symmetry handling methods have been proposed by using a rotationally invariant object model, such as polar pixels [19][20][21]25 and the generalized natural pixels. 26 Nonexact approaches, e.g., "quasisymmetry" 27 and approximate rotational symmetry, 15 have also been proposed.…”

Section: Introductionmentioning

confidence: 99%