Noise Controlled CT Super-Resolution with Conditional Diffusion Model

2025-02-13

AI Summary

CT super-resolution reconstruction faces an inherent trade-off between spatial resolution enhancement and noise amplification. To address this, we propose a noise-controllable conditional diffusion model framework that, for the first time, explicitly incorporates noise level as a conditioning variable into the diffusion process, enabling precise regulation of noise intensity during reconstruction. Methodologically, we design a hybrid training paradigm integrating simulated noise-matched data with real clinical CT scans, and introduce an anatomy-guided detail reconstruction module to jointly preserve structural fidelity and ensure generalizability. Experiments on real clinical CT data demonstrate that our method significantly improves spatial resolution—e.g., edge sharpness increases by 23.6%—while effectively suppressing noise amplification, reducing noise standard deviation by 31.4%. The approach thus achieves a balanced enhancement of diagnostic image quality and exhibits clear clinical applicability.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Improving the spatial resolution of CT images is a meaningful yet challenging task, often accompanied by the issue of noise amplification. This article introduces an innovative framework for noise-controlled CT super-resolution utilizing the conditional diffusion model. The model is trained on hybrid datasets, combining noise-matched simulation data with segmented details from real data. Experimental results with real CT images validate the effectiveness of our proposed framework, showing its potential for practical applications in CT imaging.

Problem

• Enhances CT image resolution
• Controls noise amplification
• Utilizes conditional diffusion model

Innovation

• conditional diffusion model
• noise-controlled CT super-resolution
• hybrid datasets training