A conservation approach to digital image datasets: Safeguarding cultural integrity in the digital age
Date:
Digital reproductions have become the primary means through which artworks are accessed, studied and circulated. However, many exist in uncurated and unstable forms – having been, for example, cropped, outdated, blurred, colour-modified or digitally altered – leading to distortions that affect interpretation, attribution and reuse. This paper advances a conservation approach to digital image datasets, arguing that digital surrogates should be regarded as integral components of cultural heritage rather than neutral reproductions. Drawing on a perception study and complementary computer-vision analysis, it examines how variations in image fidelity and metadata integrity influence aesthetic judgment, cultural understanding and algorithmic performance. Using WikiArt as a case study, structural inconsistencies are revealed that perpetuate visual bias and outdated representations of artworks. Finally, a framework for ethical digital curation is proposed grounded in provenance tracking, metadata consistency and transparency standards to safeguard cultural memory and promote interpretive fidelity in digital-heritage practice.
Recommended citation: De Winter, S. & Koßmann, L. (2026). A conservation approach to digital image datasets: Safeguarding cultural integrity in the digital age ICOM-CC 21st Triennial Conference, Oslo, Norway.