If you've got a scanner, chances are you've also got Optical Character Recognition (OCR) software that came packaged with it. If you've never used it, take a little time right now to get to grips with its complexities before reading on.

Back so soon? Perhaps you're already an OCR veteran, and you've scaled the dizzy heights of scanning and converting text before. Exhilarating, isn't it. But where do you go from here? You've scanned the text, you've sent the resume, you've given in to that nagging doubt an hour later and actually proof-read it ("Deep sir, pleasc fiend inclosed..."). So where to now?

Here are a few suggestions that will enable you to unleash the true power and potential of your OCR software :

  • Try OCRing a page of your own handwriting. The kind of results you get will depend on a number of factors, not the least of these being the length of time since you last actually wrote something using pen and paper. After screwing up my eyes and sticking out my tongue, I managed to scrawl the first line of this write-up onto a piece of notepaper and offer it up to the creative genius that is my OCR software; it responded in part with its own special emoticon - &~) - plus the following :

    0_k(ocki)

  • You're ready for the next stage. Time to dig out some old photos and start OCRing! Be aware that you may not get great results right away, and it may be necessary for you to prepare your material to some degree first. Basically, what you're after here are lots of small and well-delineated areas of high contrast, so contrast/brightness settings and image filters such as Photoshop's 'Find Edges', 'Glowing Edges', and 'Unsharp Mask' can help here, as can careful choice of initial material. A heavily filtered photo of tree branches produced (in part) the following :

    \ T ~ f UkK

    Yes, honestly. So you see, it really is worth the effort.

  • Now we come to the real esoterica, the point at which human and machine truly merge in the pursuit of art - the body-part OCR. I have to be honest here, this is an imprecise science and I have had very limited success with it so far, but possibly I am expecting too much too soon. My ideal is to have the OCR respond with something along the lines of :

    ~ mm- yeh

    or something even more profound, if that's possible. As with photographs, it may be necessary to do a little work on the scan before it yields results, and I suspect that the initial material is just as important. I'm thinking scarred and wrinkled here...

Log in or registerto write something here or to contact authors.