Modeling of women shoes sizing system based on 3D foot scanner result using machine learning approach

Abstract

Accurate shoe sizing plays a crucial role in ensuring comfort, performance, and consumer satisfaction, particularly for women whose foot shapes exhibit considerable anatomical variability. To address this challenge, this research proposes a data-driven modeling framework for developing a women’s shoe sizing system based on three-dimensional foot scanner data. The study was carried out through a systematic process consisting of data preprocessing, clustering using the K-Means algorithm, and evaluation of the clustering performance. The clustering analysis identified four optimal clusters within the dataset, representing distinct patterns in foot dimension measurements. The evaluation result, with a Silhouette Score of 0.25, indicates a moderate yet acceptable level of cohesion and separation among the clusters. These findings demonstrate that the proposed model can effectively capture the underlying structure of women’s foot morphology, providing a scientific foundation for establishing more accurate, customized, and ergonomically appropriate shoe sizing standards.