The concept of Optical Character Recognition has been around for a long time but it has never gained as much popularity as it promises to gain this time if Google goes on to impelement it in Google Docs.
For the benefit of those unaware, OCR tools are those that can match patterns and extract text from the images, for instance, scanned documents. So if you find a newpaper cutting, pages of books etc that you want to share with others in textual format, you no longer need to type the whole content. You can scan the document/page and upload it to Google Docs OCR, the program would convert the image into editable text!
This feature has not been fully implemented in the standard version but you can have a peek at it here
Currently, it does not handle complex images well enough (read crashes) but if the image resolution is good (character be at least 10 pixels in height) and the text inside the image was originally ‘written using Latin character sets’1, it works pretty fine. To get an early idea, you could download the image provided in the page, make some modifications, upload it and click on “Start OCR Import”. The extracted text will be available as a new document in Google Docs.
Apparently, Google Search uses OCR techniques but with ‘scanned PDF files only’1 but since Bing has already introduced Visual Search , one can expect Google Search to implement OCR features quite soon. But then what about noisy images, CAPTCHA ? Guess, somebody should come up with a tip answering that question