UniSino: Physics-Driven Foundational Model for Universal CT Sinogram Standardization

TL;DR

UniSino is a physics-driven foundational model that standardizes CT sinograms directly in the projection domain, improving robustness and generalization across diverse undersampling scenarios for better image reconstruction.

Contribution

It introduces a universal, physics-informed foundation model for CT sinogram standardization that operates in the projection domain, unlike existing image domain models.

Findings

Outperforms existing methods in sinogram enhancement quality.

Demonstrates strong generalization across multiple datasets.

Effective in both single and mixed undersampling cases.

Abstract

During raw-data acquisition in CT imaging, diverse factors can degrade the collected sinograms, with undersampling and noise leading to severe artifacts and noise in reconstructed images and compromising diagnostic accuracy. Conventional correction methods rely on manually designed algorithms or fixed empirical parameters, but these approaches often lack generalizability across heterogeneous artifact types. To address these limitations, we propose UniSino, a foundation model for universal CT sinogram standardization. Unlike existing foundational models that operate in image domain, UniSino directly standardizes data in the projection domain, which enables stronger generalization across diverse undersampling scenarios. Its training framework incorporates the physical characteristics of sinograms, enhancing generalization and enabling robust performance across multiple subtasks spanning four benchmark datasets. Experimental results demonstrate thatUniSino achieves superior reconstruction quality both single and mixed undersampling case, demonstrating exceptional robustness and generalization in sinogram enhancement for CT imaging. The code is available at: https://github.com/yqx7150/UniSino.