Foot shape classification using 3D scanning data

Abstract

3D foot anthropometric data was collected and classified into several foot shapes. A 3D foot anthropometric database which contains 1835 male right foot scanning images was used. The subjects aged from 18 to 60 years old. A 3D foot scanner was used to collect 12 foot dimensions including foot length, ball of foot length, outside ball of foot length, foot breadth, heel breadth, ball circumference, instep circumference, toe height, navicular height, instep height, toe 1 angle and toe 5 angle. The principle component analysis (PCA) and K-means cluster analysis was applied to classify male subjects' foot shapes. The PCA results indicated that foot breadth, foot length and navicular height were selected as three principle components. The percentage of total variance explained by the 3 principle components was 72.96%. The use of K-means clustering can classify male subjects' foot into 6 foot types. In addition, a new foot sizing system was developed for Taiwanese males. Comparing the new sizing system with the current CNS 4800-S1093 sizing system, the new sizing system can reduce the size numbers, and provide updated foot dimensions. Thus, the manufacturer can apply these results for shoe last design and footwear production with better fitness and lower cost.