68 страниц. 2012 год. LAP Lambert Academic Publishing In order to define a new model of foot sizes and for building the most comfortable shoes, we try to analyze dimensional characteristics of feet (lengths, perimeters, widths, heights and angles)to define new models of foot. We describe 6 methods to modelize foot size: two statistical models, two models based on neural network (model of sizes and a model of half sizes) and two others that integrate neural networks and fuzzy logic: it exploits the properties of learning of neural networks with learning functions blurred, to improve the footwear styles. we compare between these approaches to evaluate their performance and know what is the method given most models of foot size.