Nuance OmniPage is the best known name in ‘optical character recognition’ (OCR), the process of converting scanned or photographed pages into editable text documents of the type used by word processors. Version 18 of the program shows the level of accuracy and flexibility that’s now been reached.
How it works
The program can perform OCR from scanned pages, image files or photos. Photos can now include those taken with mobile phones and tablets, as long as they are of least 2 megapixels in resolution. The program offers a number of automated ‘workflows’, from one document format to another, such as PDF to Word or camera image to searchable PDF.
All offer the same basic four-step process: load the file, recognise the page, resolve any queries where the program can’t understand characters, and save the finished target document. OmniPage 18 is fast and pretty accurate – though it still does throw up a few odd queries in the test documents we tried; surprisingly easy words, which should definitely have been in its dictionary.
The new program has introduced several useful extensions, such as being able to work with files held in the cloud, taking documents from Dropbox or Evernote, for example, OCRing them and placing the recognised versions back on the same online servers.
Converting PDF files from normal to searchable, so you can find any words in them, but without altering the form of the PDF, is a neat trick. Being able to convert PDF or Word documents to speech files, for the sight-impaired, is another. With Nuance’s separate expertise in text-to-speech systems, this option does remarkably well.
- Now works from mobile phone and tablet photos.
- Not wonderful at maintaining page layouts.
For quick, easy and generally accurate OCR, OmniPage 18 is a good, well-priced choice. While we were surprised at some of the comparatively basic errors it made in conversion, they were few and required only minimal intervention. Facilities to recognise documents from photos and from online file stores increase the usefulness of the application.