FUSION OF CULTURAL HERITAGE AND AI IN SCULPTURAL REPRODUCTION

Abstract [English]

Cultural heritage and artificial intelligence have brought art reproduction to a new level of sophistication, which is data-driven and capable of rendering a sculpture with a high level of accuracy, interpretability, and preservation value. Manual production and interpretation has been the main source of traditional sculptural replication that caused inconsistency and restrictions in restoring fragmented or eroded artifacts. Recent developments of 3D scanning, photogrammetry, multispectral imaging, and archival digitization have allowed the high-resolution acquisition of geometric, textural, and material characteristics that make strong datasets to be used in computational reconstructions. In this paper, a deep learning architecture, generative, and shape-completion networks, are combined to achieve a unified framework that can be used to improve sculptural restoration workflows. AI-based approaches facilitate automated derivation of features, the identification of historical motifs and the forecastive modeling of missing or degenerated areas on the context-sensitive basis. Neural rendering also enhances the recovery of texture in order to generate material properties that are in line with the known cultural aesthetics. The hybrid digital-physical pipeline suggested in this paper will include the use of both automatic reconstructing and manual handover by artisan knowledgeable input, which will guarantee the appropriateness of the cultures and the faithfulness of the style. The technologies of additive manufacturing, robotic milling, and high-precision fabrication allow the physical implementation of the models created by AI and preserve the integrity of structures and visual quality. There is a multi-criteria fidelity assessment model which is based on geometric accuracy, visual realism, historical validity and material coherence.