Grasshopper

algorithmic modeling for Rhino

I am trying to read text from images; let's say an image has 2 or 3 words is there a way I can transform these words into text? I don't know where to start from....and if yes what should be the image extension file?

Thanks!

Views: 2954

Replies to This Discussion

You're looking for an OCR algorithm (Optical Character Recognition). This is difficult but you may be able to find a dll somewhere that does it? If you have a .NET or PInvokable dll it can be tied to Grasshopper using a scripting component.

There's certainly nothing in GH or the .NET framework that would do this.

David, there seems to be a number of .net libraries that support this ex: PUMA.NET Here's a full list: WIKIPEDIA. Unless of course i misunderstood the question. 

After about 4 different failed attempts with different libraries I was finally able to find one that did the trick. Take the attached zip file, right click it, open properties, choose "unblock," then unzip its contents to your GH libraries folder. there should be a new OCR component under the Human tab - feed it the path to an image (I tested it w/ tif and jpg, probably supports others) and it will attempt to read the text from the image. 

https://www.dropbox.com/s/e3khk7bto874s3z/HumanOCR.zip?dl=0

Andrew, thank you so much it works!! just a quick comment, I noticed it can only read text and not numbers, it if can read both, it would be perfect!

Can't thank you enough,

Dalia.

Hi Dalia. The component is currently set up to read both text and numbers. If you can upload a sample image I can try to troubleshoot the issue. 

Andrew, I was trying to play with it to find the best image properties to read, I realized the text needed to be very crisp and grayscale, however I also noticed that the text should be typed, I mean I can't read images with text that is a bit tilted or rotated (perspective like from a taken image) I don't even know if this is something possible, attached an image

Attachments:

I think this is going to be very difficult. The library I ported into grasshopper is designed for reading scanned text, not photographs. 

Hi Andrew,

This post is a bit old but it looks like you may have moved the files from this link, Is there any chance you can try sharing again? I'm really interested to see what you've done with these components.

Thanks!

Hi Andrew, your link no this.

 ¿Where can I find it?.

thank you very much!

Amazing! Any chance you could update the link??

Many Thanks!

Hi Andrew, this looks exactly like what I was looking for but the Dropbox folder is empty. Any chance you can re-share the component or guide me to how to have OCR for Rhino Grasshopper?

hi there,

you can try to start with this one:

http://www.codeproject.com/Articles/196168/Contour-Analysis-for-Ima...

Best,

Raul

RSS

About

Translate

Search

Videos

  • Add Videos
  • View All

© 2024   Created by Scott Davidson.   Powered by

Badges  |  Report an Issue  |  Terms of Service