Related Experiment Video
Updated: May 16, 2025
09:33
Diffusion Tensor Magnetic Resonance Imaging in the Analysis of Neurodegenerative Diseases
Published on: July 28, 2013
28.2K
Noise Controlled CT Super-Resolution with Conditional Diffusion Model
Yuang Wang1,2, Siyeop Yoon1, Rui Hu1
1Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Boston MA 02114, USA.
Conference Proceedings. International Conference on Image Formation in X-Ray Computed Tomography
|April 4, 2025
Summary
This study presents a novel conditional diffusion model for CT image super-resolution, effectively controlling noise amplification. The method enhances spatial resolution in CT scans using hybrid training data, proving effective in real-world applications.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Image Processing
Background:
- Improving spatial resolution in CT images is crucial but challenging.
- Noise amplification often accompanies super-resolution techniques in CT imaging.
Purpose of the Study:
- To introduce an innovative framework for noise-controlled CT super-resolution.
- To leverage conditional diffusion models for enhanced CT image quality.
Main Methods:
- Developed a conditional diffusion model for CT super-resolution.
- Trained the model on hybrid datasets: noise-matched simulations and real segmented details.
- Validated the framework using real CT images.
Main Results:
- The proposed framework effectively improves spatial resolution in CT images.
- Noise amplification is successfully controlled during the super-resolution process.
- Experimental results demonstrate the framework's effectiveness on real CT data.
Conclusions:
- The conditional diffusion model offers a promising approach for noise-controlled CT super-resolution.
- The framework shows significant potential for practical applications in CT imaging.
- This method addresses key challenges in enhancing CT image resolution and quality.
