Institutions invest tremendous amounts of time, money, technology and human capital to digitize their collections. But surprisingly, few apply even basic statistical process control (SPC) strategies to monitor the output quality of their imaging workflows. Perhaps it's the misplaced notion that digital imaging is error free or losslessly correctable. Maybe hardware manufacturers have seduced users into assuming that image quality is a foregone conclusion. Or, perhaps institutions would welcome such strategies but are simply not enabled with appropriate resources and knowledge to effectively practice them. All three of these play a role, but we believe the latter is the greatest obstacle to implementing SPC strategies.The benefits of SPC in industry are recognizable in terms of quality, efficiency, and economy. The same can occur with digital imaging in the cultural heritage community. Any SPC program naturally involves monitoring selected output parameters. The short list could be sampling rate (dpi), resolution, noise, and tonal/color fidelity. By way of a suitable target artifact and analysis software, each of these variables can be periodically measured and their values compared to pre-established numerical aims and error bounds. Corrective action is taken when trends approach or exceed these bounds. This paper describes efforts at the Library of Congress to introduce SPC practices into imaging workflows by supporting the development of unique hardware and software tools that provide ISO standardized imaging performance measurements.In addition to advocating the incorporation of genuine SPC practices into the imaging workflows of cultural heritage institutions, we also present some encouraging progress on enabling ISO imaging performance compliant tools to accomplish this in a workflow and archiving friendly fashion. Results exercising these tools are shared, and future approaches are presented. Because of its fundamental importance, the image capture stage is the focus of this paper. However, the principles involved apply equally to display, printing, and metadata generation.
Michael Stelmach, Don Williams, "When Good Scanning Goes Bad: A Case for Enabling Statistical Process Control in Image Digitizing Workflows" in Proc. IS&T Archiving 2006, 2006, pp 237 - 243, https://doi.org/10.2352/issn.2168-3204.2006.3.1.art00054