To process high-volume input data?such as the scanned images of publisher?s book and journal collections?content understanding systems should run automatically, continuously, and without human attendance. Ensuring the output quality of such systems is a challenging task, however, and automated quality assurance (QA) techniques are thus essential to its success. In this article, the author discusses three automated QA techniques that were developed for HP?s Digital Content ReMastering (DCRM) system.