Two-dimensional rigid-body kinematics using image contour registration

Area of Science:

• Biomechanics
• Medical Imaging
• Kinesiology

Background:

• Quantifying planar joint motion is crucial for diagnosing musculoskeletal conditions.
• Traditional methods often require invasive markers, limiting in vivo applications.
• Noninvasive techniques are needed for accurate and practical kinematic analysis.

Purpose of the Study:

• To present a novel method for calculating 2D rigid-body kinematic parameters using bone shape features.
• To enable noninvasive quantification of planar joint motion in vivo.
• To validate the accuracy of the proposed method.

Main Methods:

• Utilized digitized bone contour images from two positions (e.g., CT, radiographs).
• Employed best-fit contour curvature registration to map bone images.
• Generated displacement vectors to estimate planar rigid-body kinematic parameters.
• Validated accuracy using cadaveric femoral bone images under simulated pure rotation and translation.

Main Results:

• The method achieved accurate estimation of planar rigid-body kinematic parameters.
• For pure rotation, average rotation error was 0.3 ± 0.8 degrees and translation error was 0.9 ± 0.5 mm.
• For pure translation, average rotation error was -0.01 ± 0.69 degrees and translation error was -0.62 ± 0.98 mm.
• Accuracy was independent of rotation magnitude.

Conclusions:

• The developed method provides accurate 2D kinematic parameter estimation using shape features.
• This technique offers a noninvasive alternative for in vivo joint motion analysis.
• It is particularly valuable in scenarios where marker fixation is impractical or impossible.