Lower limb estimation from sparse landmarks using an articulated shape model.

Area of Science:

• Biomechanics
• Musculoskeletal Modeling
• Medical Imaging

Background:

• Accurate patient-specific musculoskeletal models are crucial for clinical gait analysis.
• Current methods using motion capture data often lack detailed bone geometry and muscle attachment information.
• Standard model scaling techniques do not account for natural anatomical variations.

Purpose of the Study:

• To develop an automated workflow for generating patient-specific lower limb musculoskeletal models.
• To improve the accuracy of bone geometry, pose, and muscle attachment site estimation from motion capture data.
• To create a more robust and realistic lower limb model for gait and force analysis.

Main Methods:

• Development of an articulated statistical shape model for the left lower limb.
• Integration of anatomical landmarks and muscle attachment regions within the shape model.
• Implementation of an automated workflow using motion capture data and a software application.

Main Results:

• The automated workflow accurately estimates lower limb bone geometry, pose, and muscle attachment regions.
• Validated bone models showed significantly lower surface-to-surface root-mean-squared error (4.28mm) compared to standard scaling (5.22mm).
• Errors at anatomical landmarks were also reduced (8.6mm vs. 10.8mm) with the new method.

Conclusions:

• The presented method provides more accurate and realistic lower limb bone geometries and muscle attachment sites.
• This approach enhances patient-specific musculoskeletal modeling for improved clinical gait analysis.
• The shape model-constrained method offers a significant improvement over traditional lower limb model scaling techniques.