The article presents a method to reconstruct 3D models of historical buildings based on archival photographs. This challenging task is crucial for preserving cultural heritage. The authors use neural rendering methods to recover the 3D shape of buildings and introduce a color appearance embedding loss to recover the color of the buildings from limited color images. The study also presents a new historical dataset of the Hungarian National Theater as a benchmark for the reconstruction method.
Publication date: 29 Nov 2023
Project Page: https://arxiv.org/abs/2311.17810v1
Paper: https://arxiv.org/pdf/2311.17810
