My Opticbook3600 software saves page images in png format. I think that or tiff is recommended for OCR purposes, as jpg is a lossy format that doesn't do well with sharp edges, if I recall correctly. It takes me about an hour to scan a 250 page book, and less than five more minutes to do the OCR. I save both so that when I come back and proof the OCR text I can look at the page image to figure out the bad conversions. I've found that ABBYY cheap edition works best with grey-scale images, not b/w - I've never felt the need to make any manual adjustments to the scanned images.
|