All geek questions in one place

Create optimized / OCR-friendly camera activity

I am trying to create camera activity for taking OCR'd photos. Here is what I want to do:

  • The resized window in the middle of the camera preview to indicate which specific area will be created in the bitmap and processed by the OCR engine.
  • Continuous autofocus (done)

I am using tesseract btw.

If someone wants to point me to some links / examples / tutorials, that would be great.

+4
source share
3 answers

I did something like that. Right now, I'm just sending the whole photo to a web service and processing it using OCRfeeder , which will segment the image and send each part with text in it for tesseract. This way I get much better accuracy. In addition, you can do some preprocessing to clear the image first.

+1
source

There may be two general approaches.

You can resize the image before sending it to the OCR engine. Keep in mind that the Tesseract mechanism you use has a kind of function - it requires some space between the characters and the borders of the images, sometimes more than expected.

The second approach is to use field-level recognition when you specify the coordinates of a text block and send the full image to the OCR engine. Take a look at http://www.ocrsdk.com , this is a cloud-based OCR SDK with web api, recently launched by ABBYY, it is in beta, so it is free to use at the moment. It has field level methods and Android code samples . I work @ABBYY and can provide additional information about our products, if necessary.

0
source

Source: https://habr.com/ru/post/1393377/

More articles:


All Articles