eduzhai > Applied Sciences > Engineering >

AHP-Net adaptive-hyper-parameter deep learning based image reconstruction method for multilevel low-dose CT

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 17 pages

Abstract: Low-dose CT (LDCT) imaging is desirable in many clinical applications toreduce X-ray radiation dose to patients. Inspired by deep learning (DL), arecent promising direction of model-based iterative reconstruction (MBIR)methods for LDCT is via optimization-unrolling DL-regularized imagereconstruction, where pre-defined image prior is replaced by learnabledata-adaptive prior. However, LDCT is clinically multilevel, since clinicalscans have different noise levels that depend of scanning site, patient size,and clinical task. Therefore, this work aims to develop anadaptive-hyper-parameter DL-based image reconstruction method (AHP-Net) thatcan handle multilevel LDCT of different noise levels. AHP-Net unrolls ahalf-quadratic splitting scheme with learnable image prior built on frameletfilter bank, and learns a network that automatically adjusts thehyper-parameters for various noise levels. As a result, AHP-Net provides asingle universal training model that can handle multilevel LDCT. Extensiveexperimental evaluations using clinical scans suggest that AHP-Net outperformedconventional MBIR techniques and state-of-the-art deep-learning-based methodsfor multilevel LDCT of different noise levels.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×