One-to-many Reconstruction of 3D Geometry of cultural Artifacts using a synthetically trained Generative Model

Pöllabauer, Thomas; Kühn, Julius; Li, Jiayi; Kuijper, Arjan

One-to-many Reconstruction of 3D Geometry of cultural Artifacts using a synthetically trained Generative Model

Files

081-084.pdf (2.77 MB)

Date

2023

Authors

Pöllabauer, Thomas
Kühn, Julius
Li, Jiayi
Kuijper, Arjan

Publisher

The Eurographics Association

Abstract

Estimating the 3D shape of an object using a single image is a difficult problem. Modern approaches achieve good results for general objects, based on real photographs, but worse results on less expressive representations such as historic sketches. Our automated approach generates a variety of detailed 3D representation from a single sketch, depicting a medieval statue, and can be guided by multi-modal inputs, such as text prompts. It relies solely on synthetic data for training, making it adoptable even in cases of only small numbers of training examples. Our solution allows domain experts such as a curators to interactively reconstruct potential appearances of lost artifacts.

CCS Concepts: Computing methodologies → Reconstruction; Supervised learning; Applied computing → Archaeology

        @inproceedings{10.2312:gch.20231161
,
booktitle = {Eurographics Workshop on Graphics and Cultural Heritage
},
editor = {Bucciero, Alberto and 
Fanini, Bruno and 
Graf, Holger and 
Pescarin, Sofia and 
Rizvic, Selma
},
title = {{One-to-many Reconstruction of 3D Geometry of cultural Artifacts using a synthetically trained Generative Model
}},
author = {Pöllabauer, Thomas and 
Kühn, Julius and 
Li, Jiayi and 
Kuijper, Arjan
},
year = {2023
},
publisher = {The Eurographics Association
},
ISSN = {2312-6124
},
ISBN = {978-3-03868-217-2
},
DOI = {10.2312/gch.20231161
}
}