X-ray Source Motion Deblurring Framework for Fast Scan in Digital Breast Tomosynthesis.

Area of Science:

• Medical Imaging
• Radiology
• Image Reconstruction

Background:

• Wide-angle digital breast tomosynthesis (DBT) acquisition is limited by X-ray source motion blur and inherent ripple artifacts.
• These artifacts can obscure subtle lesions and reduce diagnostic accuracy.

Purpose of the Study:

• To develop a deblurring technique that accelerates wide-angle DBT by reducing motion blur and accounting for ripple artifacts.
• To improve lesion visibility and preserve image texture in DBT.

Main Methods:

• A novel deblurring framework was developed, modeling the point-spread function (PSF) as a depth-dependent kernel.
• The framework includes a High-Attenuation Artifact Reduction (HAR) module for ripple suppression and a Ripple Artifact-Considered Deblurring (RAD) module.
• RAD utilizes a convolutional neural network (CNN) regularizer and analytical data-fitting to handle artifacts and restore sharpness.

Main Results:

• The proposed method successfully enhanced lesion visibility and preserved textures on both numerical and physical phantom data.
• No significant ringing artifacts were introduced, and quantitative evaluation metrics showed promising results.
• The technique effectively suppressed high-attenuation artifacts and source motion blur.

Conclusions:

• A ripple-aware deblurring pipeline enables faster wide-angle DBT by allowing higher X-ray tube speeds.
• This approach offers a practical solution to shorten patient compression time and improve clinical throughput.
• The method maintains perceptual fidelity while jointly addressing artifacts and motion blur.