A vision based proposal for classification of normal and abnormal gait using RGB camera.

Area of Science:

• Biomechanics
• Kinesiology
• Medical Diagnostics

Background:

• Human gait involves complex coordination of the entire motor system.
• Identifying gait abnormalities early can aid in diagnosing diseases.
• Current gait analysis often focuses on specific parameters, with potential for broader application.

Purpose of the Study:

• To develop a functional specification for gait analysis focusing on observational kinematic aspects.
• To create a system for classifying normal versus abnormal human gait.
• To explore the potential of gait analysis in rehabilitation, sports, and developmental pediatrics.

Main Methods:

• Development of an image sequence analysis system to extract gait features.
• Focus on lower body kinematic aspects for gait characterization.
• Identification of dynamic parameters like heel strike, toe off, stride length, and time.

Main Results:

• The prototype successfully determines dynamic gait parameters and skeleton joint information.
• The system demonstrates a satisfactory ability to distinguish between normal and abnormal gait patterns.
• Experiments were conducted on a dataset comprising 30 samples.

Conclusions:

• The proposed functional specification and extraction system offer a robust method for gait analysis.
• This approach has the potential for wide application across various fields requiring gait assessment.
• The ability to differentiate normal from abnormal gait supports early detection of potential health issues.