How to use OpenCV / Tesseract to recognise Text on a poster images

asked 2016-11-14 00:52:23 -0500

Dear All,

I am new to OpenCV and Tesseract. Currently, I am working on a project which needs to recognise text on some poster images. I tried Tesseract but the result is not very good. I am thinking of should I use OpenCV to pre-process the image before using Tesseract to recognise the text. Can anyone kindly give me some high-level directions on what kind of OpenCV technologies I should applyC:\fakepath\vip2_trim_80.jpg. I attached a sample image that I want to extract text from the image.

Thank you very much for your help.

Samuel

edit retag flag offensive close merge delete

Comments

1

Tesseract works better on an image that is

  • Rectified, so no lens or position distortion
  • A binarized edge map of the text to analyse

How to get there depends on how much variation you want to allow

StevenPuttemans gravatar imageStevenPuttemans ( 2016-11-14 06:12:50 -0500 )edit