Comparison of rarefication techniques for foot simulation using subject specific three-dimensional anthropometry data

Abstract

It is believed that one important extrinsic factor that causes foot deformity and pain was the footwear design. A good fit is one of the determinant design characteristics that determine the user’s comfort. The fitness is not only geometric match between the shoe last and a static foot but also the fit during walking or running. With the availability of subject-specific three-dimensional foot data, it would be possible to construct a foot surface model for the specific subject and to use the model as a quick and intuitive tool for evaluation of the fitness of a shoe. In this paper, we used subject-specific three-dimensional anthropometry data to build static foot surface model and applied forward kinematics into the foot model to drive it. The static model allows the fitness test of shoe for one time try on, while the dynamic model allows evaluation of the fitness during walking or other activities (with shoe model and gait information given). The anthropometry data were firstly rarefied to ensure the efficiency of data processing. The reduced dataset was further segmented into phalanges using the marker points at the joints of the foot. Finally, the foot surface model was constructed and driven phalanx by phalanx to imitate the movement of human beings. Data rarefication problem was specifically addressed in this paper since it was an art to balancing the data accuracy and the computation efficiency. In total, six rarefication techniques were compared based on three principles: the calculation speed, the visualization effect, and the volume of the specific phalanx. The arch-height technique was selected as the most suitable technique for the reduction of foot anthropometry dataset.