One CT Unified Model Training Framework to Rule All Scanning Protocols

2026-03-16

AI Summary

Existing non-ideal measurement CT (NICT) enhancement methods rely on paired data or assume homogeneous noise, limiting their generalizability across diverse scanning protocols and susceptibility to organ motion artifacts. This work reveals for the first time that different CT scanning protocols form distinct submanifolds in feature space and introduces an Uncertainty-guided Manifold Smoothing (UMS) framework. UMS employs a submanifold classifier to identify protocol types and estimate associated uncertainties, dynamically fusing global and submanifold-specific features to construct a continuous, dense feature space. Operating in an unsupervised manner, this approach dispenses with the conventional assumption of noise homogeneity and achieves significant improvements in cross-protocol reconstruction quality and model generalization across multiple public datasets.

Technology Category

• Machine Learning: Learning with Manifolds
• Reasoning under Uncertainty: Uncertainty Representations
• Multiagent Systems: Multiagent Systems under Uncertainty

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Non-ideal measurement computed tomography (NICT), which lowers radiation at the cost of image quality, is expanding the clinical use of CT. Although unified models have shown promise in NICT enhancement, most methods require paired data, which is an impractical demand due to inevitable organ motion. Unsupervised approaches attempt to overcome this limitation, but their assumption of homogeneous noise neglects the variability of scanning protocols, leading to poor generalization and potential model collapse. We further observe that distinct scanning protocols, which correspond to different physical imaging processes, produce discrete sub-manifolds in the feature space, contradicting these assumptions and limiting their effectiveness. To address this, we propose an Uncertainty-Guided Manifold Smoothing (UMS) framework to bridge the gaps between sub-manifolds. A classifier in UMS identifies sub-manifolds and predicts uncertainty scores, which guide the generation of diverse samples across the entire manifold. By leveraging the classifier's capability, UMS effectively fills the gaps between discrete sub-manifolds, and promotes a continuous and dense feature space. Due to the complexity of the global manifold, it's hard to directly model it. Therefore, we propose to dynamically incorporate the global- and sub-manifold-specific features. Specifically, we design a global- and sub-manifold-driven architecture guided by the classifier, which enables dynamic adaptation to subdomain variations. This dynamic mechanism improves the network's capacity to capture both shared and domain-specific features, thereby improving reconstruction performance. Extensive experiments on public datasets are conducted to validate the effectiveness of our method across different generation paradigms.

Problem

• Non-ideal measurement CT
• scanning protocols
• unsupervised learning
• sub-manifolds
• image reconstruction

Innovation

• Uncertainty-Guided Manifold Smoothing
• sub-manifold adaptation
• unpaired CT reconstruction
• scanning protocol generalization
• dynamic feature fusion