3D Reconstruction of Incomplete Archaeological Objects Using a Generative Adversarial Network

TL;DR

This paper presents ORGAN, a GAN-based 3D reconstruction method that effectively restores incomplete archaeological objects by predicting missing geometry, conditioned on metadata, with high accuracy even when large parts are missing.

Contribution

The paper introduces a novel GAN architecture for 3D object reconstruction that incorporates conditioning on metadata and combines multiple loss functions for improved accuracy.

Findings

Successfully reconstructs damaged objects with over 50% missing voxels

Maintains low error rates in reconstruction

Effective across diverse archaeological object types

Abstract

We introduce a data-driven approach to aid the repairing and conservation of archaeological objects: ORGAN, an object reconstruction generative adversarial network (GAN). By using an encoder-decoder 3D deep neural network on a GAN architecture, and combining two loss objectives: a completion loss and an Improved Wasserstein GAN loss, we can train a network to effectively predict the missing geometry of damaged objects. As archaeological objects can greatly differ between them, the network is conditioned on a variable, which can be a culture, a region or any metadata of the object. In our results, we show that our method can recover most of the information from damaged objects, even in cases where more than half of the voxels are missing, without producing many errors.