Developments in scanning technology

Researchers at the University of Toyko have developed a new scanner that operates at a rate of 500 frames per second which means that you can simply flip the pages of a book in front of the scanner for it to work. Overlaying a grid on the pages being scanned means that you can leave it up to the computer to flatten out the curves of the page as it was scanned resulting in an augmented, though still fairly good, scanned image. What’s next? Drive by OCR? The only problem I can see is that the most likely candidates for this kind of scanning would be in copyright material – material that is out of copyright is usually a bit older and more fragile and probably shouldn’t be subjected to the flip treatment … but it’s still an exciting development!

Take a  look at how easy the process is in this video.

Google goggles

What is Google goggles? It’s a Google developed application that utilises the camera on your mobile device and uses Google search to cross reference the image (which might be something like a landmark) to give you more information about it. It’s early days for this product but it’s an innovative step giving people the ability to query the world around them.

Currently in beta testing but close enough to being released that Google have already demonstrated the next stage of development in this product, Google goggles is moving into the world of mobile translation. Use your phone’s camera to take a picture of text you want translated, web connectivity to send the image which will be run through an OCR process, translated by the Google translate tool and the translation neatly sent back to your phone.

Whilst the current beta version is only translating German to English, it’s bringing the universal translator one step closer to reality.

Power to the people!

As some of you might be aware, the National Library of Australia has been undertaking a huge newspaper digitisation project. Whilst OCR software has come a long way in providing access to these texts, nothing is perfect. It is, in fact, users of this material that are making the outputted OCR text even better by providing feedback and correcting the OCR transcriptions as they go. The Australian Newspaper Digitisation Program states that “Users have corrected over 2 million lines of electronic text in over 100,000 articles. Over 46,000 tags have been added to articles and many comments about information in articles added.”

There is an article in the latest issue of D-Lib Magazine by Rose Holley, the Manager of the Australian Newspaper Digitisation Program, that covers this in more detail.