Automated quality assurance for document understanding systems

To process high-volume input data, such as the scanned images of publishers´ 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 techniques are thus essential to its success. The author discusses three automated QA techniques that were developed for Hewlett-Packard´s Digital Content ReMastering system.