Comparative evaluation of techniques for the harmonic analysis of human motion data

Area of Science:

• Biomechanics
• Signal Processing
• Computational Mathematics

Background:

• Accurate estimation of Fourier coefficients is essential for analyzing pseudoperiodic, non-exact discrete functions.
• Human motion photogrammetry generates complex data requiring robust analytical techniques.
• Existing methods for Fourier coefficient estimation vary in their approach and applicability.

Purpose of the Study:

• To comparatively evaluate four distinct mathematical techniques for estimating Fourier coefficients.
• To assess the suitability of these techniques for processing human motion photogrammetric data.
• To provide practical guidance for selecting appropriate methods based on specific criteria.

Main Methods:

• Numerical harmonic analysis with and without data interpolation.
• Fitting empirical data with analytical models for Fourier integral calculation.
• Comparative evaluation based on accuracy, error information, data requirements, and computational factors.

Main Results:

• Quantitative assessment of the accuracy of Fourier coefficient estimates for each technique.
• Evaluation of each method's ability to provide information on the accuracy of its estimates.
• Analysis of the a priori information and computational implementation factors for each technique.

Conclusions:

• The study provides practical insights for non-specialist users to choose the optimal Fourier coefficient estimation technique.
• Recommendations are offered based on the trade-offs between accuracy, computational cost, and data characteristics.
• This research aids in the precise analysis of human motion using photogrammetric data.