Nijs, M. (Myrna) (2020) The reliability and validity of observer based forensic gait analysis. thesis, Human Movement Sciences.
Full text available on request.Abstract
Introduction: There are many different gait patterns and a combination of features of gait can be specific to a person. Forensic gait analysis, identification based on a gait pattern, is a new and growing topic. Since a computer program is not always able to correctly quantify a gait pattern, several methods of analysis are being developed, which are based on expert human observations. The expertise of expert human observers at identifying gait features from video footage has not been tested previously, because normally the actual gait pattern is not known. Purpose: Evaluating the capability and reliability of expert human observers to correctly score gait features based on video camera footage. Method: The gait pattern of 20 healthy adult participants (11 male / 9 female; age 25.3±6.4 years) was measured during two trials of normal walking over 5 meters (approximately 7-8 steps). The gait pattern was captured with 3 high definition cameras and through a full-body motion capture system. The footage from the video cameras was scored by two expert observers. Results: The Spearman correlations between the observer scores and the motion capture scores are mostly not significant. However, the Spearman correlations between camera positions are mostly significant. Overall, the percentages of correct scoring are satisfactory or high. Cohen’s kappa and the percentage of agreement between both observers show a high interrater reliability. Conclusion: The results from this study suggest that expert forensic gait observers are able to mostly score gait features from camera footage correctly and show that the interrater reliability of this observer based identification method is high.
| Item Type: | Thesis (UNSPECIFIED) |
|---|---|
| Supervisor name: | Otten, prof. dr. E. |
| Faculty: | Medical Sciences |
| Date Deposited: | 20 May 2022 09:25 |
| Last Modified: | 20 May 2022 09:25 |
| URI: | https://umcg.studenttheses.ub.rug.nl/id/eprint/3370 |
Actions (login required)
 |
View Item |