PulseAugur
EN
LIVE 04:45:48

New neural network slashes parameters for CT reconstruction

By PulseAugur Editorial · [2 sources] ·

Researchers have developed a new Gaussian-Based Shift-Variant filtered backprojection (FBP) neural network, named GB-SVFBP, for efficient reconstruction in non-circular trajectory cone beam computed tomography. This novel approach integrates a trainable 2D Gaussian model into the filtering process, significantly reducing the number of trainable parameters by 99%. The method also accelerates convergence by reducing training time to one-fourth of the original, while only slightly impacting reconstruction quality. AI

IMPACT This research offers a more efficient method for CT reconstruction, potentially improving speed and reducing computational costs in medical imaging applications.

RANK_REASON The cluster contains a research paper detailing a novel neural network architecture and its performance improvements.

Read on arXiv cs.CV →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

New neural network slashes parameters for CT reconstruction

COVERAGE [2]

  1. arXiv cs.CV TIER_1 English(EN) · Chengze Ye, Linda-Sophie Schneider, Yipeng Sun, Andreas Maier ·

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

    arXiv:2607.11584v1 Announce Type: new 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 differenti…

  2. arXiv cs.CV TIER_1 English(EN) · Andreas Maier ·

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

    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 filte…

RELATED ENTITIES

RELATED TOPICS