A kinematic-geometric model based on ankles' depth trajectory in frontal plane for gait analysis using a single RGB-D...

Area of Science:

• Biomechanics
• Computer Vision
• Human Motion Analysis

Background:

• RGB-D cameras and pose estimation offer potential for biomechanics.
• Challenges in markerless gait analysis include noise and inaccuracy.
• Existing methods may require specific markers or complex setups.

Purpose of the Study:

• To develop a novel kinematic-geometric model for accurate spatio-temporal gait analysis.
• To overcome limitations of existing markerless gait analysis techniques.
• To enable portable and precise gait analysis using depth data.

Main Methods:

• A kinematic-geometric model based on ankle trajectories in the frontal plane and depth data.
• Modeling gait patterns using parameterized curves and a simultaneous fitting algorithm.
• Numerical optimization to minimize geometric and biomechanical error functions.

Main Results:

• The model demonstrated good to excellent statistical agreement (R_c between 0.86 and 0.99) compared to a gold standard system.
• Accurate computation of spatio-temporal gait parameters using only ankle depth data.
• The approach effectively denoises and smooths ankle depth trajectories.

Conclusions:

• The proposed model provides a robust and accurate method for markerless gait analysis.
• It relies solely on ankle depth trajectories, simplifying data acquisition.
• The model facilitates the development of portable, accurate, and markerless gait analysis systems.