Gait Analysis Clustering for Stride Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current gait analysis systems face limitations in accurately extracting individual gait profiles from diverse gait sequences, particularly in clinical settings, as they often compute stride features by averaging over heterogeneous strides, which can include various gait patterns, leading to misinterpretation of gait impairments and requiring a large number of strides for reliable results.
Innovation Solution
A method and system for analyzing gait that involves identifying and clustering stride features from 3D movement data, allowing for the definition of classes of strides, enabling the selection of representative strides and providing insights into gait patterns and impairments with a minimum number of strides, even in complex movement sequences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If stride features are computed by averaging over heterogeneous strides, then a large number of strides can be analyzed, but the accuracy of gait profile extraction deteriorates due to mixing different gait patterns
Solution Approach 1:
The patent segments the gait data by clustering strides into homogeneous groups based on similarity metrics. Instead of treating all strides as a single homogeneous set for averaging, the method divides them into distinct clusters representing different gait patterns, thereby preserving accuracy while enabling analysis of large quantities of diverse strides
Solution Approach 2:
The patent applies local quality by computing stride feature averages separately within each homogeneous cluster rather than globally across all strides. This allows the analysis to maintain high precision for each specific gait pattern type while still being able to process and compare multiple clusters, effectively addressing both quantity and accuracy requirements
2Reliability
If visual assessment is performed by physicians, then gait quality can be evaluated, but the assessment becomes subjective and time-dependent
Solution Approach 1:
The patent replaces the mechanical/subjective visual assessment system with an automated computational system that uses inertial sensors and algorithmic processing. The system objectively computes stride features and clusters them based on mathematical similarity metrics, eliminating physician subjectivity and time-dependency while maintaining or improving assessment quality
Solution Approach 2:
The patent enables the gait assessment system to be self-sufficient by automatically performing data collection, processing, clustering, and interpretation without requiring physician involvement in the actual assessment process. The system serves itself by using predefined algorithms to objectively evaluate gait patterns, making the process more convenient and consistent
3Measurement precision
If stationary gait analysis systems are used, then measurement precision improves, but location dependency and device complexity increase
Solution Approach 1:
The patent extracts the essential gait measurement functionality from complex stationary systems by using simplified inertial sensors that can be worn on the body. This extraction maintains sufficient measurement precision for clinical purposes while eliminating the need for complex laboratory equipment and specialized locations
Solution Approach 2:
The patent creates a universal measurement system using inertial sensors that can function in multiple settings (clinic, home, outdoor) and for various gait analysis purposes. The same simple sensor device provides consistent measurement precision across different locations and applications, replacing the need for location-specific stationary systems
4Reliability
If a large number of strides are required for reliable gait analysis, then measurement reliability improves, but patient burden and assessment time increase
Solution Approach 1:
The patent performs preliminary action by pre-defining clustering algorithms and similarity metrics before data analysis. This allows the system to efficiently organize and process stride data into homogeneous groups, extracting reliable gait patterns from fewer strides by focusing analysis on representative clusters rather than requiring exhaustive analysis of every individual stride
Data Source
AI summary
The present invention relates to methods for analyzing gait of a subject. In particular, the present invention relates to a method for analyzing gait of a subject, said method comprising: providing data representing the 3D-movement of a foot of said subject over time; identifying within said data first data segments that each represent of at least one stride; determining one or more stride features for each of said first data segments; and defining one or more clusters on the basis of at least one stride feature of said one or more stride features. Each of the defined clusters represents a class of strides, e.g. a class may represent the typical stride of a subject. The present invention also provides for corresponding systems that are configured to perform the methods of the present invention and the use of these systems for analyzing in assessing gait of a subject, preferably a subject suffering from a movement-impairment.