MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel
  1. Home
  2. Meeting Program
  3. All
  4. Low-Dose CT with Deep Learning Regularization Via Proximal Forward Backward Splitting

Low-Dose CT with Deep Learning Regularization Via Proximal Forward Backward Splitting

Q Ding1*, G Chen2,3, H Ji1, H Gao3, (1) National University of Singapore, Singapore, (2) Shanghai Jiao Tong University, Shanghai, China,(3) Winship Cancer Institute of Emory University, Atlanta, GA

Presentations

(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: Low dose X-ray computed tomography (LDCT) is desirable for reduced patient dose. This work develops image reconstruction methods with deep learning (DL) regularization for LDCT.

Methods: Our methods are based on unrolling of proximal forward-backward splitting (PFBS) framework with data-driven image regularization via deep neural networks. In contrast with PFBS-IR that utilizes standard data fidelity updates via iterative reconstruction (IR) method, PFBS-AIR involves preconditioned data fidelity updates that fuse analytical reconstruction (AR) method and IR in a synergistic way, i.e., fused analytical and iterative reconstruction (AIR). Instead of conventional image regularization, deep learning regularization via convolutional neural networks is utilized in PFBS-AIR to learn data-driven features and transforms.

Results: The performance of the proposed methods is evaluated in comparison with FBP (an AR method), TV (an IR method) and FBPConvNet (a DL-based image postprocessing method), using different low-dose levels. The quantitative results of peak signal to noise ratio (PSNR), root mean square error (RMSE) and structural similarity index measure (SSIM) are tabulated for quantitative evaluation of image quality. The table suggests that our method achieved superior performance for all low-dose levels. TV had larger PSNRs, smaller RMSEs and larger SSIMs than FBP method as expected. The DL-based methods improved the reconstructed results from FBP and TV, among which PFBS-AIR had the best quantitative reconstruction quality.

Conclusions: We have developed a DL-regularized image reconstruction method for LDCT, using the optimization framework of PFBS, with (A)IR for preconditioned data-fidelity update, namely PFBS-(A)IR. The preliminary results suggest PFBS-AIR had superior reconstruction quality over FBP (an AR method), TV (an IR method), FBPConvNet (a DL-based image postprocessing method), and PFBS-IR (a DL-regularized image reconstruction method), owing to the synergistic integration of analytic reconstruction, iterative reconstruction, and deep learning for LDCT.

Keywords

Not Applicable / None Entered.

Taxonomy

Not Applicable / None Entered.

Contact Email

Additional Links