Robust Text Recognition for Uzbek Language in Natural Scene Images

Abstract

Text recognition is a challenging and emerging research area in computer vision. The text which is located in images includes semantic data which is practiced in many various applications such as robot navigation, information retrieval of complex images, street signs, beneficial for visually impaired, automatic read the signboard, etc. We propose a text recognition method for Uzbek language using the Tesseract engine. We show that the model can recognize the Uzbek language, including Latin and Cyrillic alphabet text. The experimental results of the proposed method show robust performance than other methods.