Motion tracking and gait feature estimation for recognising Parkinson's disease using MS Kinect.

Area of Science:

• Biomedical Engineering
• Neurology
• Computer Science

Background:

• Gait analysis is crucial for monitoring neurological disorders like Parkinson's disease and assessing treatment efficacy.
• Microsoft (MS) Kinect sensors offer a novel approach to track 3D movements for gait feature detection.

Purpose of the Study:

• To develop and evaluate a method for detecting gait disorders using MS Kinect sensors.
• To assess the accuracy of MS Kinect in measuring gait parameters such as stride length and velocity.

Main Methods:

• Utilized digital signal processing for error rejection and feature extraction from MS Kinect data.
• Developed algorithms to estimate leg length, normalized average stride length (SL), and gait velocity (GV).
• Collected data from Parkinson's disease patients, age-matched healthy individuals, and students.

Main Results:

• Accurate leg length estimation with a mean difference of 0.004 m across all participants.
• Identified distinct gait patterns for Parkinson's patients (SL: 0.38 m, GV: 0.61 m/s) compared to controls (SL: 0.54 m, GV: 0.81 m/s).
• Achieved 97.2% accuracy in classifying gait disorders using neural networks with combined gait features.

Conclusions:

• MS Kinect sensors show significant potential for gait disorder recognition.
• This technology can serve as an affordable alternative to complex systems for gait analysis.
• The findings support the use of MS Kinect in both clinical and daily life settings for gait monitoring.