Kluge2017 - Full Pipeline#
Further Links:
A challenge to test the performance of a full gaitanalysis pipeline on laboratory 4x10m gait tests.
Comparisons are performed on the calculated means of spatial temporal parameters over the entire gait test. Reference gait parameters are provided by a marker-less motion capture system that covers the center of the walking section. The entire validation is run as a 5-fold cross-validation to allow the algorithms to optimize parameters on an independent train set.
General Information#
- Dataset
The Kluge 2017 dataset [1] (usage example, download) contains 4x10m gait tests of 20 TODO
- Sensor System
Two IMU sensors (Shimmer 3, 102.4 Hz) are attached laterally to the shoes of the participants.
- Reference System
A marker-less motion capture system (Simi Motion, 100 Hz) is used to track the foot and ankle trajectory. For this challenge, we use the trajectory of the ankle marker to calculate stride length. Heel strikes were labeled manually based on the video data.
Implementation Recommendations#
As the MoCap system does not cover the turns at the end of each 10m section algorithms should also cut of all turning strides (recommendation: turning angle > 20 deg) before calculating the mean and the variance. Otherwise, calculated results will have a considerable bias.
Notes#
The comparison here is fundamentally different that the validation performed in the original paper [1]. There, parameters were compared on a per-stride basis, while here the mean over the entire gait test is used. Hence, error values are not directly comparable.
References#
Kluge, Felix, Heiko Gaßner, Julius Hannink, Cristian Pasluosta, Jochen Klucken, and Björn M. Eskofier. “Towards Mobile Gait Analysis: Concurrent Validity and Test-Retest Reliability of an Inertial Measurement System for the Assessment of Spatio-Temporal Gait Parameters.” Sensors 17, no. 7 (July 2017): 1522. https://doi.org/10.3390/s17071522.