Marker-free Method for Reconstruction of Human ...

Digital image processing techniques were used to develop algorithms for automatic detection of human motion trajectories in this study. Based on these algorithms, the marker-free analysis system has been developed and tested in analysis of human gait. Experiments have been made on a group of ten healthy male subjects whose walking was captured by standard camera. Here, only the movements of lower limbs’ anatomical landmarks in sagittal plane were considered. Trajectories of hip, knee, ankle and fifth metatarsal head were reconstructed by this new system and used to calculate anatomical joint angles and spatio-temporal parameters. To assess the accuracy of the method, the real distances were measured by laboratory measuring-tape and compared with those reconstructed by the system. The average difference between the reconstructed and measured values was 3.83 mm with standard deviation 0.69 mm. Kinematical parameters were also compared with those obtained by commercial marker-based motion capture system SMART with no statistically significant differences. The laboratory tests shows satisfying results and the proposed method might be applicable to kinematical analysis of human motion.