A novel multi-view X-ray digital imaging stitching algorithm

Area of Science:

• Medical Imaging
• Digital Image Processing
• X-ray Technology

Background:

• Fan beam X-ray imaging often requires stitching multiple images due to detector size limitations.
• Stitching artifacts, particularly deformation, are a challenge in creating panoramic X-ray images.

Purpose of the Study:

• To introduce a novel multi-view X-ray digital imaging stitching algorithm (MVS).
• To address and resolve stitching deformation issues in fan beam X-ray imaging.
• To leverage CdZnTe photon counting linear array detectors for improved stitching.

Main Methods:

• The MVS algorithm involves four key steps: multi-view projection data acquisition, overlapping positioning, weighted fusion, and projected pixel value calculation.
• The algorithm was tested using fan beam X-ray scans of a globe and a foot, acquired with a CdZnTe detector.
• Comparative analysis was performed against three other stitching methods.

Main Results:

• The MVS algorithm achieved a stitching accuracy of 94.4%, a substantial improvement over the 50% accuracy of comparative methods.
• The algorithm successfully eliminated image distortion in the stitched images of the globe and foot.
• Significant improvements in overall image quality were observed.

Conclusions:

• A new multi-view X-ray digital imaging stitching algorithm (MVS) has been successfully developed.
• The MVS algorithm demonstrates superior performance and accuracy in X-ray image stitching.
• The study confirms the algorithm's feasibility and effectiveness for practical applications in medical imaging.