3D Gait Analysis From 2D Video Without Wearable Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for gait analysis, such as subjective evaluation by medical staff and sensor-based systems, lack objectivity and are cumbersome, affecting reliability and accessibility, especially for large-scale assessments.
Innovation Solution
A method and apparatus using a two-dimensional video analysis to estimate three-dimensional joint positions without markers or sensors, employing algorithms like Perspective-n-point (PNP) to calculate gait parameters from two-dimensional video data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor-based systems are used for gait analysis, then measurement precision is improved, but device complexity and ease of operation deteriorate due to sensor attachment requirements
Solution Approach 1:
The patent uses a camera to capture visual images of the subject's gait and creates a digital copy of the movement data without requiring physical sensors on the body. The image processing system extracts gait parameters from the visual copy, eliminating the need for sensor attachment while maintaining measurement capability.
Solution Approach 2:
The patent replaces the mechanical sensor attachment system with an optical imaging system. Instead of using accelerometers, gyroscopes, or pressure sensors that must be physically attached to the subject, the system uses cameras to capture gait information and processes these images computationally to obtain measurement data.
2Measurement precision
If multiple sensors are attached to various body areas for accurate analysis, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent uses a single camera system that can capture gait information from multiple perspectives and body regions simultaneously. The image processing algorithm analyzes various body parts (feet, knees, hips) from the captured images to extract comprehensive gait parameters, eliminating the need for multiple separate sensors positioned at different body locations.
Solution Approach 2:
Instead of attaching multiple physical sensors to different body areas, the system creates a comprehensive digital copy of the gait movement through visual imaging. The camera captures the entire gait cycle and body positioning, and computational algorithms extract information from multiple body regions from this single visual copy, reducing device complexity while maintaining analysis accuracy.
3Reliability
If sensor-based methods are used, then gait parameters can be measured objectively, but loss of time increases due to attachment and positioning procedures
Solution Approach 1:
The system performs preliminary calibration by capturing images of the subject in a standard position before the actual gait measurement. This establishes the coordinate system and camera parameters in advance, so that when the subject performs the gait task, the system can immediately process the images without time-consuming setup procedures.
Solution Approach 2:
The patent replaces the time-consuming mechanical process of sensor attachment and positioning with a non-contact optical imaging system. The camera captures gait information instantly without requiring physical contact with the subject, eliminating the time needed for sensor application, adjustment, and verification while maintaining objective measurement capability.
Data Source
AI summary
A gait analysis method includes collecting a two-dimensional gait video of a subject's gait situation during a preset time, which is captured using a camera; generating a three-dimensional coordinate system based on at least one frame of the two-dimensional gait video by receiving three-dimensional information of a capturing space; calculating three-dimensional feature point coordinates by receiving a plurality of feature points on the three-dimensional coordinate system, and calculating two-dimensional feature point coordinates for the plurality of feature points using resolution information of the camera; estimating mapping information between the three-dimensional feature point coordinates and the two-dimensional feature point coordinates; calculating a plurality of two-dimensional joint position information for the subject from the two-dimensional gait video; calculating a plurality of three-dimensional joint position information corresponding to the plurality of two-dimensional joint position information; and calculating gait parameters for the subject using the three-dimensional joint position information.