Technical Note: Correction of geometric x-ray image intensifier distortion based on digital image correlation

Area of Science:

• Medical Imaging
• Image Processing
• Digital Image Correlation

Background:

• X-ray image intensifiers (XRIIs) are prone to geometric distortions.
• Existing correction methods have limitations, including sensitivity to noise and patch discontinuity.
• Global methods are less effective for local distortions, while local methods struggle with noise and discontinuity.

Purpose of the Study:

• To introduce a novel local method for correcting geometric distortion in XRIIs.
• To address the limitations of existing global and local distortion correction techniques.
• To present a method based on digital image correlation (DIC) that is robust to noise and avoids patch discontinuity.

Main Methods:

• Digital Image Correlation (DIC) was employed as the core technique for distortion analysis.
• The method's performance was validated using simulated distortions, including noise and blurring.
• Experimental validation was conducted on a 14" XRII across four different orientations.

Main Results:

• Simulated distortion corrections achieved an average accuracy of (0.20 ± 0.04) pixels.
• Experimental data from the XRII showed distortions ranging from 4.4 ± 1.3 mm to 6.1 ± 2.5 mm on average.
• The DIC method achieved a weighted mean residual error between (0.037 ± 0.003) mm and (0.054 ± 0.003) mm, independent of XRII orientation.

Conclusions:

• The proposed DIC-based technique is unaffected by Gaussian noise and blurring.
• The method achieves accuracy comparable to or exceeding current standard tools for XRII distortion correction.
• The novel method is fast, automatable, and requires minimal operator intervention.