Related Experiment Video
Updated: Mar 22, 2026
06:54
Clinical-oriented Three-dimensional Gait Analysis Method for Evaluating Gait Disorder
Published on: March 4, 2018
14.9K
Reliable Gait Recognition Using 3D Reconstructions and Random Forests - An Anthropometric Approach
Martin Sandau1,2, Rikke V Heimbürger3, Karl E Jensen4
1Department of Neuroscience and Pharmacology, University of Copenhagen, Blegdamsvej 3, DK-2200, Copenhagen, Denmark.
Journal of Forensic Sciences
|April 29, 2016
Summary
Manual annotation of 3D gait data for forensic analysis shows promise. While consistent expert annotation ensures reliable recognition, interobserver variability impacts accuracy, highlighting the need for standardization in 3D forensic gait analysis.
Area of Science:
- Forensic Science
- Biomechanical Engineering
- Computer Vision
Background:
- Photogrammetric measurements and gait analysis from CCTV are crucial for forensic investigations.
- Accurate extraction of bodily dimensions and gait parameters from visual data presents significant challenges.
Purpose of the Study:
- To evaluate the reliability of manual annotation of joint centers on 3D reconstructions for forensic gait analysis.
- To determine if 3D reconstructions offer a viable alternative to conventional CCTV for gait pattern recognition.
Main Methods:
- Sixteen participants underwent continuous 3D reconstruction during normal walking.
- Eight expert observers manually extracted segment lengths and kinematics from the extremities.
- Interobserver reliability was assessed by comparing recognitions based on data annotated by the same versus different experts.
Main Results:
- High participant recognition rates (100%) were achieved when data was annotated by the same expert.
- Recognition accuracy decreased to 72.6% when data was annotated by different experts, indicating significant interobserver variability.
- Specific kinematic parameters were identified as being heavily affected by interobserver variability.
Conclusions:
- 3D reconstructions are feasible for forensic gait analysis, offering an improvement over traditional CCTV methods.
- Standardization of annotation protocols is necessary to mitigate interobserver variability and enhance reliability.
- Further research is required to address variations in clothing and environmental conditions for real-world forensic applications.
