CORRESPONDENCE MATCHING IN 3D MODELS FOR 3D HAND FITTING

Abstract

Upper limb prosthetic is an area of medical research and development that aims to restore functionality and improve the quality of life of people affected by the loss of one or both upper limbs. The development and implementation of 3D scanning tools and analysis of 3D scanning data requires the use of specialized analysis methods that ensure the achievement of the required indicators. It should take into account the impact of the model resolution on the result. This paper is devoted to the analysis of finding matches between a point cloud of a hand model and another point cloud using Gromov-Wasserstein distance. For analysis, a subset of the MANO dataset was employed, containing a substantial volume of data and serving as a representative sample of the human population. The results obtained indicate the possibility of using this approach in the processing and analysis of three-dimensional data, which serves as one of the stages of designing individualized prostheses.