The Influence Of Marker Set Protocols And Optimization Algorithms On The Estimation Of Reconstructing Human Lower Limbs 3D Kinematics

Objective: The object of the research is confirming the marker set protocol and the optimization algorithm that could estimate lower limbers 3D kinematics more correctly through comparing the hip angles,the knee angles and the ankle angles of same motions estimated by HH anatomical landmark protocol(Helen Hayes anatomical landmark protocol)with single-body optimization,HH anatomical landmark protocol with multi-body optimization,rigid marker cluster protocol with single-body optimization and rigid marker cluster protocol with multi-body optimization.Method : Capture the motion of level walk,jog and squats jump by HH anatomical landmark protocoland rigid marker cluster protocol simultaneously,the motion capture system is Motion Raptor-4 with 7 high-speed infrared lenses(200Hz).Estimate joint angles of motions that tracked by one marker set protocol with single-body optimization and multi-body optimization in Visual 3D.Calculate CMC(Coefficient of Multiple Correlation,is a variable that estimates the consistency of waveform of multiple curves)and RMS(Root Mean Square,represents the average difference of multiple curves or variability)of joint angle.Compere the reliability of every way estimating joint angle and compere the differences of joint angles estimated by 4 methods.Result:(1)For level walk,the reliabilities of all ways estimating the joint angles are high except HH anatomical landmark protocol with multi-body optimization(the CMCs of joint angles are above.75 and the RMSs are below 2°).The difference of joint angle between estimated by HH anatomical landmark protocol with multi-body optimization and by rigid marker cluster protocol with multi-body optimization is relatively small(for instance,the CMC of adduction/abduction angle of hip estimated by these 2 ways is.899 and the RMS is 1.86 °).The difference of joint anglebetween estimated by HH anatomical landmark protocol with single-body optimization and by HH anatomical landmark protocol with multi-body optimization is relatively small(such as the CMC of dorsiflexion/plantarflexion angle estimated by these 2 ways is.944 and the RMS is 1.20°).(2)For jog,the reliabilities of all ways estimating the joint angles are high except HH anatomical landmark protocol with multi-body optimization(the CMCs of joint angles are above.75 and the RMSs are below 3 °).The difference of joint angle between estimated by HH anatomical landmark protocol with single-body optimization and by marker cluster protocol with single-body optimization is relatively small(such as the CMC of flexion/extension angle of hip estimated by these 2 ways is.969 and the RMS is 2.07°).The difference of joint angle between estimated by HH anatomical landmark protocol with single-body optimization and by HH anatomical landmark protocol with multi-body optimization is relatively small(for example,the CMC of adduction/abduction angle of ankle estimated by these 2 ways is.847 and the RMS is.83°).(3)For squats jumps,the reliabilities of all method estimating the joint angles tend to decline(the CMC of adduction/abduction angle of hip is.440-.618,the RMS of flexion/extension angle of knee is 6.67°-7.30°).The difference of joint angle between estimated by all 4 ways tends to incline(for instance,the CMC of internal/external rotation angle of ankle estimated by these 2 ways is.648 and the RMS is 2.18°).Conclusion:(1)The compensation effect of optimization algorithms is getting bad when soft tissue artifact is becoming large.(2)Compering with HH anatomical landmark protocol,the compensation effect of optimization algorithms is relatively good while track marker protocol is rigid marker cluster protocol.(3)The error of joint angle estimated by rigig marker cluster protocol with multi-body optimization is relatively small when studying level walk using motion capture system.(4)The effect of optimization algorithms compensating soft tissue artifact of squat jumps is poor.