Improvement via Multiple Depth-Camera Motion Capture System for Biomechanics Rehabilitation Analyses

Biomechanics motion capture system has a huge turning point with the introduction of depth-camera technology. This technology is serving clinical experts, including physiotherapists, to analyze patients’ rehabilitation process related to joints and segments motions. Due to the simplicity of its setup, three-dimensional analysis capability, and marker-less data acquisition, this technology has been proposed as a potential replacement for the conventional two-dimensional optical system. On the other side of its simplicity, depth-camera marker-less motion capture devices have frequently been subjected to accuracy issues relative to the conventional systems. While having duplex depth-camera devices has been shown to have some improvement in accuracy by some researchers, this work demonstrated a clear sight of how accurate a depth-camera motion capture system can be if multiple depth-camera devices are used. A novel method of measuring the system's accuracy was used where the segmental length of the arm was measured in hundreds of motion frames and compared with the actual segmental lengths. Both average and root-mean-square error (RMSE) throughout the data showed significant, almost 20%, improvement in terms of accuracy compared to dual-camera depth-camera systems. Since the duplex depth-camera device system had already been proven to be more accurate than the monocular depth-camera device system, this work has further confirmed the ever-important hypothesis where the accuracy of depth-camera motion capture system can be enhanced by increasing the number of the depth-cameras used. The current findings also indicated that marker-less multiple depth-camera motion capture systems could better replace the conventional industrial-standard optical biomechanics motion capture system.