An iterative projection-based motion estimation and compensation scheme for head x-ray CT.

Area of Science:

• Medical Imaging
• Radiology
• Image Reconstruction

Background:

• Patient motion during computed tomography (CT) scans can lead to artifacts, necessitating repeat scans or sedation.
• Current CT technology aims for rapid head scans, but motion artifacts remain a challenge.

Purpose of the Study:

• To develop and evaluate a method for estimating and compensating patient motion during CT image reconstruction.
• To improve the quality of head CT images affected by motion artifacts.

Main Methods:

• An iterative approach was used to estimate and compensate rigid motion during reconstruction.
• Motion was estimated view-by-view and used to update the system matrix in each iteration.
• A multiresolution scheme accelerated the joint estimation of image and motion, followed by a final iterative reconstruction.

Main Results:

• The method significantly improved reconstructed image quality in simulations, phantom studies, and patient scans.
• Root-mean-square error decreased, and mean structural similarity increased in simulation and phantom data.
• Motion blurring was substantially reduced in patient CT images after applying the compensation technique.

Conclusions:

• The proposed method effectively eliminates motion-induced artifacts in head CT scans.
• The technique can be retrospectively applied to clinical helical CT scans using only measured raw data.