Prior Image Constrained Compressed Sensing (PICCS) and Applications in X-ray Computed Tomography

Abstract

A new image reconstruction algorithm, prior image constrained compressed sensing (PICCS), will be reviewed in this paper. PICCS enables accurate image reconstruction with high contrast-to-noise ratio from undersampled projection data sets. Several clinically relevant applications are reviewed to demonstrate how the new algorithm can be utilized to: reduce radiation dose, provide high quality four dimensional cone beam CT images used for image guided radiation therapy, achieve high temporal resolution cardiac cone beam CT for image guided cardiac interventions, enable perfusion measurements with micro CT and significantly improve temporal resolution in diagnostic multi-detector cardiac CT. The computational speed concerns for this iterative algorithm are also discussed and a method to accelerate the reconstruction using commercially available graphic cards is presented. Future research directions using the PICCS algorithm are also briefly discussed.

Interventional 4-D C-Arm CT Perfusion Imaging Using Interleaved Scanning and Partial Reconstruction Interpolation

“…For example, prior image constrained compressed sensing (PICCS) is a technique to reconstruct time-resolved images [35]- [37]. In CT brain perfusion imaging, PICCS was investigated in order to reconstruct images using only a few projections (e.g., [4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20] distributed over the full view-angle range and thus saving X-ray dose [38].…”

“…A further application of PICCS is to improve temporal resolution by using limited sector reconstruction (TRI-PICCS) [37], [40]- [42]. The projections are taken from a shorter viewangle range (e.g., 100…”

A Survey on Different Types of CT Image Reconstruction

“…According to this principle, reducing the total variation of the signal subject to it being a close match to the original signal, removes unwanted detail whilst preserving important details such as edges. Stimulated by the theory of compressive sampling or compressive sensing (CS) [1], [2], the sparsity based computed tomography (CT) has been a hot topic for various applications such as dose reduction [3]. Because the x-ray decrease coefficient often varies gently within an anatomical component, and large changes are usually confine around borders of anatomical structures, the discrete gradient transform (DGT), a set of finite difference operators, has been widely utilized as a sparsifying action in CS-inspired CT reconstruction such as in, whose L1-norm is also referred to as the total variation (TV) [4], and the equivalent reconstruction techniques are called TV minimization.…”

Sparse-view neutron CT reconstruction of irradiated fuel assembly using total variation minimization with Poisson statistics

“…Several TV based methods and their variants, including anisotropic TV [24,25], edge preserving TV [18], directional TV [12], and weighted TV [26,27] were proposed to improve the performance of the algorithm. Prior image constrained compressed sensing (PICCS) [6,28] and their variants [29,30] were introduced for dynamic and perfusion CT where a prior high-quality reconstruction is utilized as a constraint in the reconstruction process. The difference between the prior high-quality image and the current estimate serves an additional sparsified transform that enabled significant improvements in the undersampling factors.…”