Digitizer in my pocket: Google Docs Android app updated with OCR
Well, this is an interesting concept…although I’m not too satisfied with the output yet.
I have an Android phone (a Samsung Captivate).
Google just updated their Google Docs app, to include the ability to take a picture of a document with the phone, upload to Google Docs, and have it convert words in the picture to text.
Hypothetically, you could pick up a book, snap a picture of each page, and convert it to text…which you could then put on your Kindle.
Sounds kind of James Bond, doesn’t it?
Well, I haven’t practiced much with the camera yet, and I tried a few snaps without getting great results.
First, there was an issue of which way to hold the phone (camera). I got more than one shot where it looked right-side up in the phone, but was sideways (and not converted into text at all) in the doc. Sending the photo worked fine, by the way.
Second the OCR (Optical Character Recognition) just wasn’t that accurate…although again, that may improve based on my camera skills.
I got a decent shot of the back of my DragonNaturally Speaking box. This is a small excerpt for journalistic purposes, so it should be covered under Fair Use:
“Step Tyisìmg, Start TallfángDragon’ NaturallySpeaking“ Home makes it faster and easier to create and edit documents or without typing. Just say words and they appear on the screen.Say It and Your PC Does It
Outside of that, though, it did work.
If I can get a whole shot of a page of text at once, this could be very useful. I’d only use it with public domain materials or those for which I have the rights.
I could also see it being used for documents someone brings to a meeting…right into Google Docs for everybody to see and edit. The picture shows up too, by the way, but you can delete that.
To take the picture, you open up the Docs app, and there is an icon up by the search magnifying glass.
Tap that, and with your phone in portrait mode (taller than it is wide), you’ll see a choice for “Document from photo”.
Tap that, and take your picture.
If you like it, click Save.
You can rename it on the next page, verify your account, and click OK.
It will show up in your Google Docs online (if you already have an account, of course).
From there, just open it up. The converted text will be below the picture.
You can also open it right on the phone…just tap the document it. There’ is a choice on the single document screen to edit it.
This may help you appreciate how you can have so many errors in OCR’d docs…it just isn’t that accurate.
Some of you are probably a lot more skilled with a cell phone camera than I am. If you give it a try and let me know how it works, I’d appreciate that.
This post by Bufo Calvin originally appeared in the I Love My Kindle blog.