Showing posts with label OCR. Show all posts
Showing posts with label OCR. Show all posts

Friday, May 29, 2015

OCR shoot-out: Microsoft OneNote versus Google Docs

I wanted to copy the text from the photo of a document.  Using my phone, I uploaded the JPG file to a OneNote page using Microsoft Office Lens and then in the Modern version of OneNote, I selected the image and choose "Copy Text".  This is what I got:

Teacher Is  
She is reullg good  
She me about  
She alwags sags  
This is Teacher  
Over Ibe summer she will  
L give my teacher q:  
(Che I  
n thumbs up Mg favorite thing  
gears old.  
n Five Stars  
(D Love  
Teacher Karen Is  
She Is reallg good  
gears old.  
She me about  
She alwags sags  
This is Teacher  
Over the summer she will  
L give teacher q:  
n An A for Awesome A +  
n Two thumbs up Mg favorite thing about her is  
n Five Stqrs  
CD Love 


Quite horrible.  So I uploaded the image to Google Drive from my desktop, right-clicked on the file in Google Drive > Open with > Google Docs.  Result:

Shetாghtme bout
I give M teqcher :


Shetாghtme bout
I give :
| | Two thumbs | lawrite thing abouther is
Five Stars O

Even worse.  In fact, I had to rotate the photo to be portrait for Google Docs to spit anything at all out.  So for this particular photo, OneNote "wins".

Monday, April 11, 2011

Convenient OCR solution

Back in the bad old days, each scanner you used came with some software to do Optical Character Recognition and it was a hassle to become familiar with each software, especially when you upgraded operating systems/changed computers and lost the install disks!  This article, Top 5 Free OCR Software Tools To Convert Images Into Text, talks about some other software, but I recently discovered that Google Docs does a good job of performing OCR.  Simply:

  1. Scan document that you want to convert to text, rather than re-typing 
  2. Upload scanned image to your Google Docs, making sure to check the option "Convert text from PDF or image files to Google Docs documents."
  3. The converted image will retain the same name as your original image file, but it will be a Google word processor document with the image embedded at the beginning followed by the converted text.  Results are very acceptable although some massaging will be necessary.
Another benefit to using Google Docs rather than some of the free online OCR web sites out there is the security/privacy in case your document is sensitive (you just have to trust Google!).