Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Cultural Heritage 3D Reconstruction with Diffusion Networks

Pablo Jaramillo, Ivan Sipiran

TL;DR

This article explores the use of recent generative AI algorithms for repairing cultural heritage objects, leveraging a conditional diffusion model designed to reconstruct 3D point clouds effectively, and demonstrates significant potential in artifact restoration research.

Abstract

This article explores the use of recent generative AI algorithms for repairing cultural heritage objects, leveraging a conditional diffusion model designed to reconstruct 3D point clouds effectively. Our study evaluates the model's performance across general and cultural heritage-specific settings. Results indicate that, with considerations for object variability, the diffusion model can accurately reproduce cultural heritage geometries. Despite encountering challenges like data diversity and outlier sensitivity, the model demonstrates significant potential in artifact restoration research. This work lays groundwork for advancing restoration methodologies for ancient artifacts using AI technologies.

Cultural Heritage 3D Reconstruction with Diffusion Networks

TL;DR

This article explores the use of recent generative AI algorithms for repairing cultural heritage objects, leveraging a conditional diffusion model designed to reconstruct 3D point clouds effectively, and demonstrates significant potential in artifact restoration research.

Abstract

This article explores the use of recent generative AI algorithms for repairing cultural heritage objects, leveraging a conditional diffusion model designed to reconstruct 3D point clouds effectively. Our study evaluates the model's performance across general and cultural heritage-specific settings. Results indicate that, with considerations for object variability, the diffusion model can accurately reproduce cultural heritage geometries. Despite encountering challenges like data diversity and outlier sensitivity, the model demonstrates significant potential in artifact restoration research. This work lays groundwork for advancing restoration methodologies for ancient artifacts using AI technologies.

Paper Structure

This paper contains 14 sections, 7 equations, 10 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Background
  4. Completion with Diffusion Models
  5. Data
  6. Generalization Experiment
  7. Evaluating Results
  8. Reconstructing Archaeological Artifacts
  9. Dataset creation
  10. Data Augmentation
  11. Training
  12. Results and comparison against DRDAP
  13. Limitations
  14. Conclusions

Figures (10)

  • Figure 1: Examples of repaired pots. Top row displays the input, while the bottom row shows the composite of the input with the generated repair. All repairs have their according distance factors.
  • Figure 2: Examples of repaired cups. Top row displays the input, while the bottom row shows the composite of the input with the generated repair. All repairs have their according distance factors.
  • Figure 3: Examples of repaired vases. Top row displays the input, while the bottom row shows the composite of the input with the generated repair. All repairs have their according distance factors.
  • Figure 4: Examples of repaired bottles. Top row displays the input, while the bottom row shows the composite of the input with the generated repair. All repairs have their according distance factors.
  • Figure 5: Bar plot of the CDF results presented in Table \ref{['table:pcdiffvdrdap']} separated by data class of the compiled dataset. Dashed bars represent 1 standard deviation in both directions.
  • ...and 5 more figures