Accessible Digital Reconstruction and Mechanical Prediction of 3D-Printed Prosthetics

Abstract

We identify two significant issues that render prosthetics inaccessible. First, obtaining a representation of the residual limb can be inaccessible. Conventional approaches require equipment or expertise often unavailable in resource-constrained communities. Second, it is challenging to determine the prosthetic design, filament material, and printing process that satisfies mechanical functionality requirements because it is difficult to predict the mechanical properties of 3D-printed prosthetics. Therefore, we propose a method to achieve a digital residual limb model from a smartphone video and predict the mechanical functionality of the 3D-printed prosthetic. We also present a case study that demonstrates the feasibility of the method. Digital reconstruction results show that the smartphone type influences reconstruction time and mesh quality, with correlation coefficients of 0.89 and 0.88, respectively. Also, the distance between the residual limb and the smartphone influences the reconstruction scale, with a correlation coefficient of –0.90. Seven of eight digital reconstruction results achieved an average deviation lower than 2 mm, which is viable for designing prosthetics. The XGBoost model trained to predict the effective material data of the 3D-printed part achieved an R2 over 0.99 for all predictions. The predicted effective material data are used to predict the mechanical functionality of a 3D-printed prosthetic. The mechanical functionality is evaluated following ISO-10328. The results reveal that different prosthetic designs, filament materials, and printing processes yield different mechanical functionality. These factors can be determined according to the predicted functionalities and the patient’s needs.