The field of OCR, or Optical Character Recognition, has more than its fair share of acronyms floating around, but one of the most potentially exciting is ADRT.

The so-called Adaptive Document Recognition Technology of Abbyy FineReader 9.0 bucks the trend of most OCR programs, aiming to consider an entire document as a whole, instead of processing single pages individually.

Quick and simple

It works a treat with ADF (Auto Document Feed) scanners but is just as capable with regular flatbed scanners, so long as you don't mind feeding in successive pages manually.

The trick to FineReader's new holistic approach is that it's designed to identify all the salient points of a document's structure, including entities such as headers, footers, page numbering, columns, tables and graphics, as well as good old-fashioned blocks of body text.

Other OCR packages do this as well but the new edition of FineReader does a better job of rebuilding documents using native formatting commands from Word or Excel.

It's a big time saver because once the OCR has been carried out, documents are pretty much ready to go, rather than requiring a bunch of random frames and text boxes to be reformatted on the hoof.

Added accuracy

ABBYY claims a rise in accuracy levels of 35 per cent for character recognition and 32 per cent for layout retention, compared with the previous edition of FineReader.

Our tests bore this out, although version 9 still isn't foolproof, occasionally misinterpreting some tables and graphics we threw at it.

Also noticeably better is the main OCR interface, which is more task-focused and offers one-click access to common jobs like converting scans, PDF files or even photos to Word or Excel files.

And for the finishing touch, output options include no less than five flavours of PDF files, including fully searchable PDFs, along with more workaday Office 2007 file format options.