CoRe-GAN: A Gated Attention Framework for High-Fidelity Restoration of the Codex Borgia

Lopez Garcia, Jose Manuel (2025) CoRe-GAN: A Gated Attention Framework for High-Fidelity Restoration of the Codex Borgia. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (2MB) | Preview
Preview	PDF (Configuration Manual) Download (735kB) | Preview

Abstract

This project presents CoRe-GAN, a gated-attention generative model for restoring damaged regions of the Codex Borgia, a highly detailed Mesoamerican manuscript. A custom dataset of segmented codex imagery was created, along with a CodexDamageMaskGenerator that simulates realistic deterioration such as cracks, pigment loss, and abrasion. CoRe-GAN combines gated convolutions, contextual attention, and multi-stage training to address the manuscript’s complex iconography. Quantitative and qualitative evaluations show that the model significantly outperforms baseline inpainting methods in PSNR, SSIM, LPIPS, and texture fidelity. The results demonstrate the value of domain-adapted AI for cultural-heritage restoration.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Garg, Mohit UNSPECIFIED
Subjects:	N Fine Arts > N Visual arts (General) For photography, see TR Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence
Divisions:	School of Computing > Master of Science in Artificial Intelligence
Depositing User:	Ciara O'Brien
Date Deposited:	02 Sep 2026 10:04
Last Modified:	02 Sep 2026 10:04
URI:	https://norma.ncirl.ie/id/eprint/9762

Actions (login required)

View Item