Abstract
Marker-based, IMU-based (6-axis IMU), and smartphone-based (OpenCap) motion capture methods are commonly used for motion analysis. The accuracy and reliability of these methods are crucial for applications in rehabilitation and sports training. This study compares the accuracy and inter-operator reliability of inverse kinematics (IK) solutions obtained from these methods, aiming to assist researchers in selecting the most appropriate system. For most lower limb inverse kinematics during walking motion, the IMU-based method and OpenCap show comparable accuracy to marker-based methods. The IMU-based method demonstrates higher accuracy in knee angle (5.74 ± 0.80 versus 7.36 ± 3.14 deg, with p = 0.020) and ankle angle (7.47 ± 3.91 versus 8.20 ± 3.00 deg, with p = 0.011), while OpenCap shows higher accuracy than IMU in pelvis tilt (5.49 ± 2.22 versus 4.28 ± 1.47 deg, with p = 0.013), hip adduction (6.10 ± 1.35 versus 4.06 ± 0.78 deg, with p = 0.019) and hip rotation (6.09 ± 1.74 versus 4.82 ± 2.30 deg, with p = 0.009). The inter-operator reliability of the marker-based method and the IMU-based method shows no significant differences in most motions except for hip adduction (evaluated by the intraclass correlation coefficient-ICC, 0.910 versus 0.511, with p = 0.016). In conclusion, for measuring lower-limb kinematics, wearable sensors (6-axis IMUs) achieve comparable accuracy and reliability to the gold standard, marker-based motion capture method, with lower equipment requirements and fewer movement constraints during data acquisition.