Saturday, May 8, 2010

Use Google Docs as Free OCR - Extract Text from Images

Hi All

I have recently written about Free OCR - Extract Text from Images. In the continuation with that post, I am here to tell you about a nice trick with Google Documents that does the same work we have learnt on extracting text from images. As you are aware, OCR is very much helpful for extracting or getting text from images, scanned papers, print-outs, wallpapers, banners, etc. I came to know this trick with Google Docs, when I was surfing online curiously about 'OCR'. I would like to share my experience with you.



Google Documents List Data API
As part of Google Labs (An innovative environment where Google tests the new features to be added in Google Services like GMail, Orkut, Google Reader, Google Calender, Google Docs), Google introduced List Data API v3.0 for its free service Google Documents. API stands for Application Programming Interface, an interface implemented by a software or a service through which one application can interact with other. 

Google Docs as Free OCR
As I told you earlier, List Data API is for extensions and innovations on Google Documents. Most recently, Google has added a new extension, Optical Character Recognition (OCR) interface to this API, which helps us to extract text from images. In the next few topics, I will clearly demonstrate how we are going to do this job.

Google Docs OCR Demo
Pre-requirement for this trick to work is that you should have your Google Documents account associated with your Free Google Account. You need an image or banner with good resolution, from which you are going to extract text. This is the link that I found from DocList API for OCR Demo: http://googlecodesamples.com/docs/php/ocr.php 
Open this link in a new tab. Sign in to your Google Docs account.



When the OCR API prompts for the access to Google Docs, click 'Grant Access' button and wait for next screen to load.



In this screen, you can find the details about OCR, features and restrictions with that. Please follow the below simple steps which are self-explanatory:

  1. Click 'Choose File' to select an image file to be scanned for text. (.JPEG, .JPG, .PNG, .BMP)
  2. Choose the image file from your hard drive.
  3. Click 'Open'
  4. Click 'Start OCR import'
  5. If you don't find any scanned file or image containing some text, but wish to test this trick, you can download a sample image file at the bottom of the page and then do the above steps.



Here is the input image file I used for the demonstration. Actually I took a screen-shot from one of my previous posts: Top 5 IMs for Mobile Phones.



Here is the output of the OCR import or scan. The text failed to extract during this import is marked as 'unrecognized text'. This is quite awesome, right? It is really useful to test this out. By default, the imported text will be saved a Google Document (A text file with the name 'the file name of the image supplied') in Google Docs.



Issues with Google Docs OCR
Google DocList OCR API is still as the Lab feature of Google Documents. So it could not be error free. If you face any issues when performing OCR import (like one given below), please check out your Google Documents - All Items, the imported file may be available there.





More About Google Docs OCR
Access your Google Docs here. Read more about Google DocList API from Google Labs API page. Read more about OCR from Wikipedia
I hope that the information is useful. Please comment about this trick.

About This Post
This post is helpful on the following searches:

  • Read text from images
  • Optical Character Recognition
  • Optical Text Recognition
  • Extract characters and text from image files
  • Image processing
  • Google Documents tips and tricks - OCR

-- Muthuraj L

No comments:
Write comments