A simple method to choose the most representative stride and detect outliers.

Area of Science:

• Biomechanics
• Medical Engineering
• Data Analysis

Background:

• Gait analysis relies on kinematic data, often presented as joint angle curves over the gait cycle.
• Repeated measurements are typical but require methods to identify representative data and outliers.
• Existing statistical methods may not fully capture curve shape and position variations.

Purpose of the Study:

• To develop an automatic, computationally inexpensive method for selecting representative kinematic curves in gait analysis.
• To introduce a technique for effectively detecting outliers among repeated gait measurements.
• To provide a robust tool for analyzing both single and multi-kinematic variables.

Main Methods:

• A novel 'depth-based' method is proposed, identifying the 'deepest' curve as the median equivalent.
• The method analyzes both the shape and position of kinematic curves.
• It is applicable to single kinematic variables and multi-variable gait profiles.

Main Results:

• The depth-based method successfully identifies the most representative gait curves.
• Outliers within repeated kinematic measurements are effectively detected.
• The approach demonstrates sensitivity to both curve shape and positional variations.

Conclusions:

• The proposed depth-based method offers an efficient and accurate approach to gait data analysis.
• This technique enhances the reliability of kinematic data by selecting representative curves and identifying outliers.
• The method provides a valuable tool for clinical and research applications in gait analysis.