Reliability-Aware CT Image Reconstruction Based on Convolution–Attention Feature Uncertainty

Read full text Summarise with VIRAT
Authors: Liam O'Connor, Sophia Martinez, Daniel Kim
Cited by: 0
References: 17

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Beyond reconstruction accuracy, reliability assessment is important for clinical adoption of learned CT reconstruction methods. Motivated by convolution–self-attention denoising models such as CTLformer, this work studies reliability-aware CT reconstruction by associating reconstructed images with uncertainty estimates. The model outputs both reconstructed images and pixel-wise uncertainty maps, allowing identification of regions with reduced confidence. Experiments on two clinical CT datasets with 50,000 slices show that uncertainty-aware reconstruction maintains comparable PSNR and SSIM to deterministic baselines while enabling effective detection of reconstruction errors. Regions with high uncertainty correspond closely to metal artifacts and low-dose noise patterns.

Fields of Study & Subjects

Medical Imaging Techniques and Applications Image and Signal Denoising Methods Advanced X-ray and CT Imaging

Publication details

Journal / source
Global Media and Social Sciences Research Journal
Volume
7
Issue
1
Pages
194–199
Publication date
20 February 2026
Language
EN