An Accurate Multi-Lingual Character Recognition Approach

Character recognition is a popular and demanded area in artificial intelligence, specifically in computer vision. In this paper, we proposed a simple and high accurate method to recognize and distinguish single characters. After some preprocessing, we divide the image of the character into four blocks, then define descriptors for each block. At last, we maintain these descriptors as templates. The accuracy of this method is high and can be used for other languages.

Keywords:

OCR , feature vector , moment , standard deviation