Image Reconstruction From Limited-Angular-Range Data Acquired with An Offset-Detector
Presentations
PO-GePV-I-19 (Sunday, 7/10/2022) [Eastern Time (GMT-4)]
ePoster Forums
Purpose: To develop and evaluate the directional-total-variation (DTV) algorithm for image reconstruction from truncated limited-angular-range (LAR) data acquired with an offset-detector.
Methods: In this work, LAR scanning configurations with an offset-detector were considered, which consist of two circular arcs with angular ranges of α₁ and α₂, separated by 90⁰. The projection data were generated from a numerical head-phantom for different combinations of angular ranges α₁ and α₂. Due to the detector-offset, the projection data were truncated on one side of the detector. The image reconstruction was formulated as a solution to a constrained optimization program in which a weighted data-l₂ fidelity is minimized under image-DTV constraints, and a DTV algorithm was developed and tailored to reconstruct images from truncated LAR data. We then evaluated the performance of DTV reconstructions in terms of quantitative metrics, including normalized root mean square error (nRMSE) and Pearson correlation coefficient (PCC).
Results: We conducted the study for image reconstruction for two-orthogonal-arc scanning configurations with LARs (α₁, α₂) = (75⁰, 75⁰), (67.5⁰, 67.5⁰), (60⁰, 60⁰), as well as single-arc scanning configurations with LARs (α₁, α₂) = (150⁰, 0⁰), (135⁰, 0⁰), (120⁰, 0⁰) by use of the DTV algorithm. Results show that LAR artifacts in the FBP reconstructions are significantly reduced in the DTV reconstructions for both two-orthogonal-arc and single-arc configurations. Moreover, due to the combination of data truncation and LAR, there still exists some artifacts in single-arc reconstructions. However, two-orthogonal-arc reconstructions are visually free of the artifacts. The quantitative results show that two-orthogonal-arc configurations yield images with minimum artifacts, which is consistent with our visual inspection.
Conclusion: We developed and tailored a DTV algorithm for image reconstruction from truncated LAR data acquired with an offset-detector. The study results demonstrate that, from truncated LAR data, the DTV algorithm can yield images with significantly reduced artifacts.
Funding Support, Disclosures, and Conflict of Interest: XP serves as the Editor-in-Chief of IEEE Transaction on Biomedical Engineering, and is a shareholder and co-founder of Clarix Imaging Co. and XPIM LLC. There is no conflict of interest.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email