Development of a Robust, Simple, and Affordable Human Gait Analysis System Using Bottom-Up Pose Estimation With a Smartphone Camera

Frontiers in Physiology, 2022 · DOI: 10.3389/fphys.2021.784865 · Published: January 5, 2022

Simple Explanation

This study introduces a new system for analyzing human walking patterns (gait) using a smartphone camera and computer vision techniques. The goal is to create a system that is affordable, easy to use, and accurate, addressing the limitations of existing methods. The system, named OMGait, uses a smartphone camera to capture videos of people walking. It then uses a computer vision algorithm called OpenPose to estimate the position of body joints and calculate joint angles during walking. The OMGait system was tested on healthy volunteers under different lighting and clothing conditions, and its accuracy was compared to standard gait analysis methods. The results showed that OMGait can measure joint angles with reasonable accuracy, even under challenging conditions.

Study Duration

Not specified

Participants

16 healthy volunteers (12 male and 4 female)

Evidence Level

Not specified

Key Findings

1
OMGait can measure hip, knee, and ankle joint kinematics with relatively good accuracy using a common mobile phone camera and a personal computer using OpenPose.
2
Unlike MS Kinect, kinematic measurements done using OpenPose are tolerant to variations in ambient lighting and the type/kind of dress worn by the subjects.
3
OpenPose algorithms can accurately detect the pose of the person of interest from multi-person background images/videos and in extremely dark and bright conditions.

Research Summary

The study presents a new gait analysis system (OMGait) that uses a smartphone camera and the OpenPose algorithm to measure gait kinematics. OMGait was found to be accurate in measuring hip, knee, and ankle joint angles, even under different lighting and clothing conditions. The proposed system overcomes limitations of conventional systems and markerless systems like Kinect.

Practical Implications

Affordable Gait Analysis

OMGait offers a cost-effective alternative to expensive laboratory-based gait analysis systems, making it accessible to small clinics and developing countries.

Non-Intrusive Monitoring

The markerless approach eliminates the need for attaching sensors or markers to the patient's body, improving comfort and ease of use.

Versatile Application

The system's tolerance to different lighting and clothing conditions expands its applicability in real-world scenarios, including settings where patients wear traditional garments.

Study Limitations

1
OMGait shows large deviations at the start and end of the gait cycle
2
it is not validated by comparing the data from the same subjects with the benchmark systems such as VICON
3
The error seen in the results is predominantly due to the systematic error in the proposed system and not the ambient conditions.

Your Feedback

Was this summary helpful?

Back to Bioinformatics