pdfocr adds an OCR text layer to scanned PDF files, allowing them to be searched. It currently depends on Ruby 1.8.7 or above, and uses ocropus, cuneiform, or tesseract for performing OCR.
To use, run:
pdfocr -i input.pdf -o output.pdf
For more details, see the manpage.
pdfocr requires tesseract and hocr2pdf. These can be provided by installing the packages tesseract-ocr, tesseract-ocr-eng (or other languages you need), and exactimage from your distribution.
pdfocr was written by Geza Kovacs
pdfocr is hosted at http://github.com/gkovacs/pdfocr
Christian Pietsch added tesseract support.