Artur Kulmukhametov, Markus Plangg, and Christoph Becker
Automated Quality Assurance for Migration of Born-digital Images
In: Archiving Conference, Archiving 2014 Final Program and Proceedings, pp. 73-78(6)
Publisher: Society for Imaging Science and Technology
Abstract:
Migration to standardized formats is a common approach for the preservation of digital objects. To ensure the authenticity of the resulting artefacts and the validity of the migration, quality assurance is essential. For large-scale migration, automated quality assurance processes are an essential prerequisite. This paper focuses on the migration processes of born-digital photographs. We describe the particular requirements for successful automation of quality assurance. A key aspect of this is the authenticity of the image, the fidelity of the rendering as it appears to an expert viewer. Automation requires us to substitute the human expert viewer with a software algorithm. The key question is whether existing image comparison mechanisms can be applied. To address it, we introduce a publicly available automated workflow relying on perceptual quality assurance measures and present an experiment testing the correlation of the automated measures to human perception.
Download: Link