METHOD FOR GROUPING 3D MODELS TO CLASSIFY CONSTITUTION

摘要

Provided is a three-dimensional model classification method of classifying constitutions. The method includes correcting color values of a frontal image and one or more profile images to allow a color value of a reference color table in the images to equal a predetermined reference color value, through obtaining the frontal image and one or more profile images of a subject including the reference color table by a camera, the reference color table including one or more sub color regions, generating a three-dimensional geometric model of the subject by extracting feature point information from the frontal image and the profile image, matching the corresponding feature point information to extract spatial depth information, after removing the reference color table region from the frontal image and the profile image, and classifying a group of the three-dimensional geometric model of the subject by selecting a reference three-dimensional geometric model having a smallest sum of spatial displacements from the three-dimensional geometric model of the subject from a plurality of reference three-dimensional geometric models stored in the database and setting the group which the selected reference three-dimensional geometric model represents as the group where the three-dimensional geometric model of the subject belongs.