| Resumen |
In this work, an algorithm towards automatic feet measurements extraction for personalized footwear design is presented. The algorithm performs orientation, localization and semantic segmentation of a point cloud produced by a 3D feet scan system. The algorithm input instance is a point cloud from a 3D scan of a user standing on their feet on a flat surface. The outputs are two point clouds; each one is a segmentation of the user´s left and right foot with its own coordinate system. The algorithm implementation is done with the point cloud library Open3D. It combines linear programming and techniques from Machine Learning such as RANSAC and DBSCAN together with the point cloud processing algorithm Screened Surface Poisson Reconstruction to process the localization and segmentation of the foot. The proposed method contributes to the extraction of anthropometric measures of the feet in order to be able to build personalized footwear, which is ongoing research. |
