Limited-view/sparse-angle computed tomography software for efficient CT reconstruction

RT&L FOCUS AREA(S): General Warfighting Requirements (GWR) TECHNOLOGY AREA(S): Sensors OBJECTIVE: Limited-view/sparse-angle CT software reducing time/dose required for 3D reconstruction compared to the filtered back projection. Reconstruction quality within industry standards & compatibility with image parameters for X-radiography & neutron radiography. DESCRIPTION: The Armaments Center's Radiography Laboratory is used to determine in a nondestructive way a product's actual internal construction and composition as compared to the specification with which its construction parameters were set. The results of this nondestructive evaluation are then used to confirm a product's quality and integrity (safety, lethality, ruggedness and survivability), and to verify claims of operational limits and prepare assets for reliable field operations. Computed Tomography (CT) is a computational method in nondestructive evaluation that produces a 3D reconstruction from a series of 2D images. To produce a CT, the item under inspection is placed between an X-ray source and detector. The object (or in some cases the source-detector pair) is rotated for a series of image acquisitions at varying angles. The collected images are supplied to software that uses one of several mathematical methods to convert the 2D images into a 3D reconstruction. In the context of nondestructive testing, X-rays and neutrons are most commonly used in the context of industrial manufacturing. Filtered back projection (FBP) is the most common method for CT reconstruction. This method produces accurate 3D reconstructions using minimal computing power at the cost of increased projections. In medical imaging, the Hounsfield scale is used to measure a material's opacity to X-ray radiation, where a higher value indicates more opacity. Muscle tissue has a Hounsfield unit (HU) value of 10-40 HU, whereas steel has a value of 20,000 HU. An FBP CT of industrial manufactured parts could require from several hundred to thousands of projections, necessitating acquisition times on the order of tens of minutes (usually 15-45 minutes.). For this reason, CT is restricted to R&D environments and is only used as a last resort in production processes. CT data provides complete knowledge of the position of relevant indications within a manufactured part. In the medical field, faster acquisition times are desired to reduce the dose delivered to a patient while industry is driven not by dose but by acquisition times while maintaining stringent requirements for contrast and resolution. Limited-view CT promises a reduced overall time by minimizing X-ray time at the cost of modestly increasing computing time. Current advances in CT algorithms make such a reconstruction method possible. The literature indicates limited view CT is a developing method that has successfully been applied case by case. Certain items possess density gradients that make them inaccessible to X-rays, demanding interrogation by neutrons. Limited-view CT can be particularly helpful in neutron radiography, where acquisition time per image can be considerably longer than X-rays. In addition to our current x-ray CT capability, our laboratory is on schedule to gain an in-house neutron radiography capability in a few years, and this installation will most likely evolve into a neutron radiography CT capability. Since no off-the-shelf software package allows for reliable, limited-view CT reconstruction, this project aims to produce the first such package for use with items relevant to the Army and apply it to both X-ray CT and neutron radiography CT. PHASE I: Develop the mathematical and software methods for limited-view/sparse-angle CT. The possibility of a vast reduction in image acquisition requirements while maintaining sufficient image quality should be theoretically established. PHASE II: Develop and demonstrate prototype limited-view/sparse-angle CT software that successfully reconstructs an item of sufficient complexity while meeting quality and acquisition objectives. PHASE III DUAL USE APPLICATIONS: The US Army can use this system for faster X-ray CTs and more robust neutron radiography CTs. Industry could use this system to increase speed of nondestructive evaluation, increasing total production. This system could also be used in the medical field for high-fidelity CTs of patients while delivering minimal X-ray doses. REFERENCES: 1. Hu, Z., Gao, J., Zhang, N. et al. An improved statistical iterative algorithm for sparse-view and limited-angle CT image reconstruction. Sci Rep 7, 10747 (2017) 2. G.A. Jones, P. Huthwaite. Limited view X-ray tomography for dimensional measurements. NDT & E International, Volume 93 Pages 98-109, ISSN 0963-8695 (2018) 3. Hu, Z. et al. Image reconstruction from few-view CT data by gradient-domain dictionary learning. J Xray Sci Technol 24, 627 638 (2016) 4. Freeman, Timothy. The Mathematics of Medical Imaging: A Beginner's Guide. Springer Texts in Mathematics and Technology 2nd Edition (2015) KEYWORDS: Nondestructive Evaluation (NDE); Computed Tomography (CT); Limited-View/Sparse-Angle CT; X-Ray; Neutron Radiography (NR); CT Reconstruction