Skip to Main content Skip to Navigation
Book sections

"Art" in X-ray tomography: Image noise reduction

Abstract : Reduction of the object dose by reducing X-ray exposure has the inevitable consequence of increasing statistical noise in the projections. A set of projections with a 10% noise, collected during the test experiment at the Institute of Crystallography RAS, were used to reconstruct a water phantom. Two different reconstruction approaches (Algebraic Reconstruction Technique (ART) and Filtered Back Projections (FBP)) were implemented. The reconstructed images also had about 10 % noise in both cases. Median filtering within each ART iterative step and averaging over images, updated and preserved during the final iteration made it possible to lower the image noise to 3%. For ART calculations, the RegART software package developed by the authors was used.
Complete list of metadatas

https://hal-insu.archives-ouvertes.fr/insu-00403085
Contributor : Pascale Talour <>
Submitted on : Thursday, July 9, 2009 - 11:15:51 AM
Last modification on : Friday, September 25, 2020 - 3:13:49 AM

Identifiers

  • HAL Id : insu-00403085, version 1

Collections

Citation

Marina Chukalina, Dmitry Nikolaev, Alexandre Simionovici. "Art" in X-ray tomography: Image noise reduction. ZELINKA Ivan ; OPLATKOVA Zuzana ; ORSONI Alessandra. Simulations in United Europe, ECMS, pp.826, 2007. ⟨insu-00403085⟩

Share

Metrics

Record views

240