A method for zinger artifact reduction in high-energy x-ray computed tomography

• A filter for zinger removal is proposed and compared with a standard filter.
• It is demonstrated that high energy x-rays are the primary cause of zinger noise.
• Simulated noise and projections are used to have an absolute truth in filter analysis.
• Greater true positives, lower false positives, and lower false negatives are achieved.
• The new method yields less error in filtered radiographs and the reconstructed volume.

In x-ray imaging, the interaction of stray x-rays with an imaging sensor may produce speckle-type noise (i.e. zingers) which produces artifacts in tomography reconstructions. Zinger reduction can be accomplished through detector design and image filtering. A standard method for zinger filtering exists but produces many false positives and high discrepancy with the true signal. An approach is proposed through selective temporal frame averaging in which zinger noise is removed with improvements over the standard method. The accuracy of the proposed approach is compared with a standard approach using simulated projections of a phantom containing known artificial zinger and locations density.