Reconstruction of 3D Foot Model from Video Captured Using Smartphone Camera

Abstract

In this paper, a method is proposed to reconstruct 3D foot model using video captured by smart phone. Scale-Invariant Feature Transformation followed by feature matching were adopted to generate consistent tracks. These tracks were further processed to generate 3D points in terms of Structure-from-motion (SFM). The accuracy of the proposed method was evaluated by comparing the difference between the reconstructed foot and the scanned foot of an artificial socked foot. The mean error between them was 0.15 mm (SD=1.42 mm). Almost 80% data had an error less than 1.5 mm. The experimental results indicated that a mapped 3D foot model could be reconstructed by this simple and efficient method, and the accuracy of this method is enough for foot measurement and footwear design.