Sparse-view CT reconstruction based on group-based sparse representation using weighted guided image filtering

Area of Science:

• Medical Imaging
• Computational Imaging
• Image Reconstruction

Background:

• Traditional guided image filtering (GIF) methods using total variation (TV) struggle with accurate reconstruction of complex clinical image details.
• Limitations in existing methods necessitate improved techniques for sparse-view computed tomography (CT).

Purpose of the Study:

• To develop a novel sparse-view CT reconstruction method.
• To improve the accuracy and detail preservation in reconstructing complex clinical images.

Main Methods:

• A new method employing group-based sparse representation (GSR) for guidance image reconstruction.
• Utilizing weighted guided image filtering (WGIF) to transfer features from the guidance image to the SART reconstruction method.
• Iterative algorithm where GSR results serve as guidance images in each step.

Main Results:

• The proposed method demonstrated superior visual quality in tests with 64 projection views.
• Achieved a Peak Signal-to-Noise Ratio (PSNR) of 48.82 for a shoulder case, outperforming existing methods.
• Experimental results confirmed enhanced performance compared to other reconstruction techniques.

Conclusions:

• The developed method effectively preserves anatomical structures in CT images.
• The technique shows significant capability in noise suppression and artifact reduction.
• The proposed weighted guided image filtering approach offers a more effective solution for sparse-view CT reconstruction.