I have played with OCR software on the iPhone off and on for many years now. I install the software, try to use it, get disappointed and remove the apps to make room for other apps.
A new version of OCRTool came out last week that uses a few of the iOS4 features. Running in the background is a key one. No longer do you snap a shot of a page of text and then have to sit and watch the OCR app work it’s way through the page. If you use OCRTool, you take the picture and move onto other things you need to get done, then come back later to get the text.
It still isn’t perfect, but that isn’t a reflection on the software or the iPhone… desktop software has many of the same issues. After having used it a bit more aggressively while doing this review it came to me that only have to clean up a little bit of text is WAY faster than having to type out all the text you needed to reference later. That includes copy/pasting into a email or other documents.
Snap a picture of any text… I generally have fun trying to make sure I don’t end up shadowing part of the page or too much reflective lighting. It might be time to break out my Ponoko stand.
After you take the picture, you are able to adjust it a bit to fit the page edge-to-edge. As well you can do a bit of clean up from a quick snap. You can see the color has been zero’d out to attempt to get maximum difference in black to white.
‘Done’ causes the OCR engine to run… which is the step that can take some time so with the new version of the software you can move off to another app while the process continues in the background. Your picture is kept on the screen above the OCR text. You can see there is a larger percent of text correct than not means it is only a matter of cleaning up some text rather than typing everything new.
Once you have the text as you like, there is a pretty good list of one-button options of what you do with that text. Copying the text to the clipboard allows you to insert it into another document you may be working on. As well, sharing the text with others can be done as text in the email, a jpeg image or as a PDF attachment.
Since this image is much more perfect than one dependent on my steady hand and lighting, the OCR gets a lot more of the text correct. All of the usual actions can be taken with the resulting text I ran through above.
A nice feature is the built in translation. Now, you can snap a picture of a magazine article you can not read (due to not knowing the language), clean it up with OCR and then convert it from/to MANY languages.
The text from your image is shown after the OCR engine has run in the language it was and in the translated language. As a reminder, the from and to language country flags are shown at the bottom of the page. So… while I speak a bit Italian, I’m not great at reading it yet… how does this look?