Individual Foot Shape Modeling from 2D Dimensions Based on Template and FFD

Abstract

A novel method to generate an individual 3D foot shape from 2D dimensions is proposed. A standard foot shape (template) was first defined as the average shape from foot data samples of 50 participants. The standard foot was sectioned at each 1% foot length so that there were 99 sections. The captured images by digital cameras were used, the bottom view of the foot can be processed to give the foot outline while a side-view gives the foot profile. The points of each section were then scaled based on known 2D information with respect the outline and profile. The other points in standard foot were global deformed with method based on direct manipulation of free form deformation(FFD) with multiple point constraints. So the individual foot shape could be created by modifying the standard foot using the measured height and width obtained from the foot outline and profile respectively. After the shape prediction and alignment, the prediction error was calculated using the dimensional difference between the predicted shape and the actual foot shape. The results show that the method proposed in this paper is robust, efficient and relatively cheaper than using a 3D scanner to determine the 3D foot shape. This research may be used to develop custom lasts for the manufacture of custom footwear without actually scanning a person’s feet.