PulseAugur
LIVE 12:28:51
research · [1 source] ·
0
research

Researchers propose residual noise learning to improve PET image denoising

By PulseAugur Editorial · Summary by gemini-2.5-flash-lite from 1 source

Researchers have developed a new framework for denoising low-dose Positron Emission Tomography (PET) images, addressing the challenge of models failing to generalize across different noise levels. Their approach, termed residual noise learning, focuses on estimating noise directly from low-dose images rather than predicting full-dose images. Experiments on large-scale datasets showed this method outperforms existing AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

Read on arXiv cs.CV →

COVERAGE [1]

  1. arXiv cs.CV TIER_1 · Yichao Liu, Zongru Shao, Yueyang Teng, Junwen Guo ·

    Rethinking Cross-Dose PET Denoising: Mitigating Averaging Effects via Residual Noise Learning

    arXiv:2604.16925v2 Announce Type: replace Abstract: Cross-dose denoising for low-dose positron emission tomography (LDPET) has been proposed to address the limited generalization of models trained at a single noise level. In practice, neural networks trained on a specific dose le…