Low-Dose CT Imaging Using a Regularization-Enhanced Efficient Diffusion Probabilistic Model.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Image Processing

Background:

• Low-dose CT (LDCT) reduces radiation exposure but increases image noise, hindering clinical tasks.
• Current denoising methods are often too slow for real-time clinical use.

Purpose of the Study:

• To develop a rapid and high-fidelity denoising framework for LDCT.
• To enhance image quality and preserve anatomical details for improved diagnosis.

Main Methods:

• Proposed a regularization-enhanced efficient diffusion probabilistic model (RE-EDPM).
• Incorporated residual guidance and hybrid perceptual/total variation regularization.
• Utilized a Swin-based U-Net and a composite loss function for noise suppression and detail preservation.

Main Results:

• Achieved high performance on public LDCT benchmarks (e.g., SSIM 0.879 for chest, 0.971 for abdomen).
• Demonstrated superior noise reduction and structural fidelity compared to state-of-the-art methods.
• Processed images rapidly (approx. 0.125s per slice), enabling near-real-time application.

Conclusions:

• RE-EDPM provides effective LDCT denoising with minimal inference time.
• Offers a balance between noise reduction and anatomical preservation for clinical deployment.
• Suitable for real-time medical imaging and enhancing low-quality images.