System-independent material classification through X-ray attenuation decomposition from spectral X-ray CT

Doniyor Jumanazarov, Jakeoung Koo, Matteo Busi, Henning Friis Poulsen, Ulrik Lund Olsen, Mihai Iovea*

*Corresponding author for this work

Research output: Contribution to journalJournal articleResearchpeer-review

451 Downloads (Pure)


We present a method for material classifications in spectral X-ray Computed Tomography (SCT) taking advantage of energy-resolving 2D detectors to simultaneously extract the energy dependence of a material's linear attenuation coefficient (LAC). The method employs an attenuation decomposition presented by Alvarez et al., and estimates system-independent material properties of electron density (ρe) and effective atomic number (Zeff), independent of the scanner, from the energy-dependent LAC measurements. The method uses a spectral correction algorithm and the energy range is truncated to exclude bins with photon starvation and spectral distortion present even after correction of detector response. A novel technique of energy bin selection is used for optimized classification performance. The method is tested against another SCT classification method called SRZE for inspecting materials in the range of 6≤Zeff≤23. Our method aims at an increase in the speed of pot processing workflow after the data acquisition, and it achieves explicitly up to 32 times better time efficiency for the reconstruction with comparable accuracy for a range of materials important in threat detection.
Original languageEnglish
Article number102336
JournalNDT and E International
Number of pages12
Publication statusPublished - 2020


  • Spectral X-ray CT
  • Electron density
  • Effective atomic number
  • Photon counting detectors
  • Security screening


Dive into the research topics of 'System-independent material classification through X-ray attenuation decomposition from spectral X-ray CT'. Together they form a unique fingerprint.

Cite this