John Cxxxx It does an incredible job recognizing text, tables, and even graphics in, for example, PDF documents containing scanned images. I recently used a trial to OCR a PDF containing the scanned pages of an old technical document (5 pages of portuguese text, including tables, images and equations, plus the original scan was skewed by about 15 degrees) and Abbyy FineReader was not only able to correctly identify the text, tables and several of the images (producing a DOCX file with the reuslt), I estimate the overall document was around 90% accurate, despite the very low quality of the original.
As a comparison I tried 7 different online (free) OCR scanners, most of which were only able to extract text, and whose accuracy was not much better than 50-60%. In the one case where a DOCX file was produced, it had only 2 of the 5 pages from the original, and only 10% of the text!
