Registration of sliding objects using direction dependent B-splines decomposition.

Area of Science:

• Medical imaging
• Computational anatomy
• Image processing

Background:

• Sliding motion in deformable image registration causes discontinuities, complicating accurate anatomical alignment.
• Existing methods struggle to precisely model and correct for sliding motion, impacting registration reliability.

Purpose of the Study:

• To develop and evaluate a novel deformable image registration method capable of handling sliding motion.
• To improve the accuracy and robustness of image registration in the presence of complex tissue deformations.

Main Methods:

• A new method utilizing multiple B-spline transforms to decompose deformation into sliding regions.
• Discontinuities are allowed at region interfaces while enforcing interface matching to prevent unrealistic deformations.
• The method was validated on 16 lung cancer patient datasets.

Main Results:

• The proposed method achieved a significantly lower target registration error (TRE) of 1.5 mm compared to a single B-spline approach (3.7 mm).
• Registration accuracy was comparable to a multi-B-spline approach without sliding constraints (1.4 mm TRE).
• The method demonstrated 37% fewer gaps and overlaps along region interfaces compared to the unconstrained multi-B-spline method.

Conclusions:

• The proposed multiple B-spline transform method effectively addresses sliding motion in deformable image registration.
• This approach enhances registration accuracy and reduces artifacts, particularly in challenging anatomical regions.
• The findings suggest improved clinical applicability for image-guided interventions and treatment planning.