GB-SVFBP: Gaussian-Based Shift-Variant FBP neural network

Ye C, Schneider LS, Sun Y, Maier A (2025)

Publication Language: English

Publication Type: Conference contribution

Publication year: 2025

Event location: Shanghai, China

DOI: 10.48550/arXiv.2607.11584

Abstract

This paper proposes a Gaussian-Based Shift-Variant filtered backprojection (FBP) neural network, which is designed for the efficient reconstruction of non-circular trajectory cone beam computed tomography. The traditional differentiable shift-variant FBP model consists of a filtering component and a backprojection process. The filtering component includes operations such as weightings, differentiations, a 2D Radon transform, and a 2D backprojection. The proposed methods build on this framework by introducing a trainable 2D Gaussian model to represent the trajectory-related part in the filtering process, achieving a substantial reduction in the number of trainable parameters. Experimental results demonstrate that the proposed model reduces the parameter count by 99%, while only sacrificing a slight amount of reconstruction quality. Furthermore, the training time for each trajectory is reduced to one-fourth of the original, significantly accelerating convergence. These enhancements demonstrate a considerable augmentation in the model's practicality and effectiveness, making it a valuable asset for real-world applications.

Authors with CRIS profile

How to cite

APA:

Ye, C., Schneider, L.-S., Sun, Y., & Maier, A. (2025). GB-SVFBP: Gaussian-Based Shift-Variant FBP neural network. In Proceedings of the 18th International Meeting on Fully Three-Dimensional Image Reconstruction in Radiology and Nuclear Medicine (Fully3D). Shanghai, China, CN.

MLA:

Ye, Chengze, et al. "GB-SVFBP: Gaussian-Based Shift-Variant FBP neural network." Proceedings of the 18th International Meeting on Fully Three-Dimensional Image Reconstruction in Radiology and Nuclear Medicine (Fully3D), Shanghai, China 2025.

BibTeX: Download