This repository contains all my work in computer vision. Currently, I'm working on an OCR (Object Character Recongition) text detection and recognition in natural scene images. The context I'm working on involves images that contain text- but not an ensemble of words. The text consists of for example "4586-H1". So, it can contain numbers and letters. Introduction
My first trial was using End to End Text Recognition code in C++ provided by OPENCV. More details can be find in their documentation: http://docs.opencv.org/3.0-beta/modules/text/doc/erfilter.html
I didn't get good results with it, so I tried other methods.
Note: I will update this file with more details. Meanwhile you can check my website blog posts about the project: http://www.sarahmestiri.com/index.php/2017/09/28/from-opencv-and-…sion-challenge-2/