The document discusses digitization workflows for enhancing and segmenting documents for optical character recognition (OCR). It describes steps for image enhancement including border removal, page curl removal, and correction of arbitrary warping. It then discusses standalone methods for segmenting text lines, words, and characters without relying on character recognition. These include a hybrid text line segmenter and density-based word segmenter that have been evaluated on historical documents with promising results. The techniques allow digitization of documents with non-standard words or layouts.