DiffNR: Diffusion-Enhanced Neural Representation Optimization for Sparse-View 3D Tomographic Reconstruction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Su, Shiyan, Zha, Ruyi, Shi, Danli, Li, Hongdong, Cheng, Xuelian
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917431110270976
author Su, Shiyan
Zha, Ruyi
Shi, Danli
Li, Hongdong
Cheng, Xuelian
author_facet Su, Shiyan
Zha, Ruyi
Shi, Danli
Li, Hongdong
Cheng, Xuelian
contents Neural representations (NRs), such as neural fields and 3D Gaussians, effectively model volumetric data in computed tomography (CT) but suffer from severe artifacts under sparse-view settings. To address this, we propose DiffNR, a novel framework that enhances NR optimization with diffusion priors. At its core is SliceFixer, a single-step diffusion model designed to correct artifacts in degraded slices. We integrate specialized conditioning layers into the network and develop tailored data curation strategies to support model finetuning. During reconstruction, SliceFixer periodically generates pseudo-reference volumes, providing auxiliary 3D perceptual supervision to fix underconstrained regions. Compared to prior methods that embed CT solvers into time-consuming iterative denoising, our repair-and-augment strategy avoids frequent diffusion model queries, leading to better runtime performance. Extensive experiments show that DiffNR improves PSNR by 3.99 dB on average, generalizes well across domains, and maintains efficient optimization.
format Preprint
id arxiv_https___arxiv_org_abs_2604_21518
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle DiffNR: Diffusion-Enhanced Neural Representation Optimization for Sparse-View 3D Tomographic Reconstruction
Su, Shiyan
Zha, Ruyi
Shi, Danli
Li, Hongdong
Cheng, Xuelian
Image and Video Processing
Computer Vision and Pattern Recognition
Neural representations (NRs), such as neural fields and 3D Gaussians, effectively model volumetric data in computed tomography (CT) but suffer from severe artifacts under sparse-view settings. To address this, we propose DiffNR, a novel framework that enhances NR optimization with diffusion priors. At its core is SliceFixer, a single-step diffusion model designed to correct artifacts in degraded slices. We integrate specialized conditioning layers into the network and develop tailored data curation strategies to support model finetuning. During reconstruction, SliceFixer periodically generates pseudo-reference volumes, providing auxiliary 3D perceptual supervision to fix underconstrained regions. Compared to prior methods that embed CT solvers into time-consuming iterative denoising, our repair-and-augment strategy avoids frequent diffusion model queries, leading to better runtime performance. Extensive experiments show that DiffNR improves PSNR by 3.99 dB on average, generalizes well across domains, and maintains efficient optimization.
title DiffNR: Diffusion-Enhanced Neural Representation Optimization for Sparse-View 3D Tomographic Reconstruction
topic Image and Video Processing
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.21518