3D visualization of damaged statues using Gaussian splatting and web interface integration
摘要
Digital heritage conservation relies significantly on advanced digital documentation techniques like photogrammetry and laser scanning. These non-invasive technologies enable precise documentation, analysis, and preservation of cultural heritage. This research presents a framework for visualizing the restoration of statues through five stages: (1) Dataset Collection, (2) Applying Several Techniques for 3D Reconstruction, (3) Selecting the Optimal Method for 3D Reconstruction, (4) Visualizing the Statue, and (5) Deploying the Statue on a Web-Based Platform. Three methods for 3D Reconstruction were compared to determine the most effective. The evaluation compared photogrammetry, Gaussian Splatting, and an integrated laser scanning–photogrammetry method. Gaussian Splatting achieves competitive geometric accuracy according to the measured Hausdorff Distances. Gaussian Splatting is considered the most suitable for rapid and resource-efficient visualization, particularly when minimal computational power and straightforward web deployment are required. The framework is implemented on an Egyptian statue of a Ptolemaic queen to illustrate its use.