eduzhai > Applied Sciences > Engineering >

X-ray Photon-Counting Data Correction through Deep Learning

  • king
  • (0) Download
  • 20210505
  • Save

... pages left unread,continue reading

Document pages: 21 pages

Abstract: X-ray photon-counting detectors (PCDs) are drawing an increasing attention inrecent years due to their low noise and energy discrimination capabilities. Theenergy spectral dimension associated with PCDs potentially brings greatbenefits such as for material decomposition, beam hardening and metal artifactreduction, as well as low-dose CT imaging. However, X-ray PCDs are currentlylimited by several technical issues, particularly charge splitting (includingcharge sharing and K-shell fluorescence re-absorption or escaping) and pulsepile-up effects which distort the energy spectrum and compromise the dataquality. Correction of raw PCD measurements with hardware improvement andanalytic modeling is rather expensive and complicated. Hence, here we proposeda deep neural network based PCD data correction approach which directly mapsimperfect data to the ideal data in the supervised learning mode. In this work,we first establish a complete simulation model incorporating the chargesplitting and pulse pile-up effects. The simulated PCD data and the groundtruth counterparts are then fed to a specially designed deep adversarialnetwork for PCD data correction. Next, the trained network is used to correctseparately generated PCD data. The test results demonstrate that the trainednetwork successfully recovers the ideal spectrum from the distorted measurementwithin $ pm6 $ relative error. Significant data and image fidelityimprovements are clearly observed in both projection and reconstructiondomains.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×