A markerless sub-pixel motion estimation technique to reconstruct kinematics and estimate the centre of mass in posturography.

A novel method to evaluate postural sway is presented: balance strategies are identified by applying a markerless image processing algorithm to video sequences obtained from commercially available systems. The motion estimation technique for the analysis of video sequences is a coarse to fine procedure based on the block matching algorithm (BMA). The method makes it possible to estimate the movement of selected elements on the scene with a sub-pixel precision. It has been done by applying a bicubic spline interpolation to the coarse results obtained by the BMA. Results achieved through the analysis of synthetic video tests make it possible to determine the accuracy of the proposed sub-pixel algorithm. Figures show how the proposed method can be confidently applied to evaluate postural sway. The proposed method has been applied to videos recorded during orthostatic posture trials in different conditions, and to combined tests where signals from a balance plate have been acquired simultaneously to the video sequence. The results show the usefulness of the proposed approach in order to evaluate balance strategies in posturography.