> If you color-scan a printed document at low-enough resolution, will it corrupt the color-printer dot patterns or are they resilient to that?
It should be completely trivial to destroy any markings in the PDF. Black and white text in particular once scanned can just be thresholded and passed through filters. Heck it seems like a trivial application of deep learning to obfuscate text. For example, it would be interesting to train a transformer to learn degradation from multiple print/scan cycles using the same source using different printers and scanners so that it learns how to fake all sorts of imprecision.
It should be completely trivial to destroy any markings in the PDF. Black and white text in particular once scanned can just be thresholded and passed through filters. Heck it seems like a trivial application of deep learning to obfuscate text. For example, it would be interesting to train a transformer to learn degradation from multiple print/scan cycles using the same source using different printers and scanners so that it learns how to fake all sorts of imprecision.