Optimizing knee biomechanical parameters during the stance phase of running using the DFIS system: a pilot study

Background
Regardless of whether it involves skin displacement (marker-based) or the identification of human body shape (marker-less), it is unavoidable to encounter kinematic errors resulting from data gathering. Therefore, this paper examines utilizing skeletal displacement captured by DFIS as an optimization technique, employing the acquired skeletal displacement as a standard to modify the motion trajectory of Vicon markers. The objective of this study was to examine the impact of both the original and modified markers displacements during the knee joint’s running stance phase on the resultant kinematics and dynamics, and through the mLSTM reduce the marker-based error induced by soft tissue artifacts.

Methods
This study recruited 12 male participants, with an average age of 25.4 ± 3.05 years, height of 178.62 ± 6.13 cm, and body weight of 75.82 ± 7.28 kg. To record the right rearfoot running stance phase, the participants ran at speeds of 5 m/s.

Results
There were significant differences found in the knee joint angle, velocity, moment, power and force between the Vicon and the Modified values. Additionally, the results show that there were significant differences in the error indicator results between the Vicon-DFIS (Modified)-Prediction by MAPE, RMSE, NRMSE in the three anatomical planes during the running stance phase.

Conclusion
The unique strategy introduced in this work may significantly reduce errors induced by soft tissue artifacts. The approach used in this study can significantly enhance the precision of sports biomechanical data evaluation, outlining the direction for future research in this domain.