A promising limited angular computed tomography reconstruction via segmentation based regional enhancement and total...

Area of Science:

• Industrial imaging
• Non-destructive testing
• Image reconstruction

Background:

• X-ray computed tomography (CT) is vital for industrial non-destructive testing.
• Artifacts in CT images arise from object size, absorption, limited views, or penetration issues.
• Existing methods like total variation (TV) minimization partially reduce artifacts but struggle with accuracy due to data deficiencies.

Purpose of the Study:

• To develop an efficient iterative algorithm for limited-angle CT reconstruction.
• To enhance true image values and suppress artifacts in reconstructed images.
• To improve the accuracy and quality of CT images in challenging scenarios.

Main Methods:

• Developed an iterative algorithm combining regional enhancement and TV minimization for limited-angle reconstruction.
• Introduced a segmentation approach to create a support mask, guiding image reconstruction.
• Incorporated a new regularization term with support knowledge into the objective function.
• Solved the optimization model using variable splitting and the alternating direction method.
• Implemented a compensation approach to refine segmentation and reduce errors.

Main Results:

• The proposed algorithm significantly suppresses artifacts in limited-angle CT reconstructions.
• Achieved higher accuracy in image reconstruction compared to other methods in simulations and real data.
• Demonstrated the ability to produce high-quality reconstructed images despite data deficiencies.
• Validated through comparative studies with simulation and real CT datasets.

Conclusions:

• The developed algorithm effectively addresses artifact issues in limited-angle CT reconstruction.
• It provides a robust solution for obtaining accurate and high-quality images in industrial non-destructive testing.
• The method shows significant improvements over existing reconstruction techniques for data-limited scenarios.