Publication Date: 2023/02/08
Abstract: Recently, image-based text extraction has becomea prominent and hard study subject in computer vision. In this article, the texts written in Kazakh are classified based on factors such as writing style and diversity of writing, and a text recognition system based on correctly defined terms is developed.Text matching is accomplished by using the EasyOCR library to the input picture to extract the areas containing the words. The words in the text are then determined using the locations acquired. To initialize the text and recognize the defined text, twodistinct functions are utilized. As a consequence, a method for recognizing Kazakh words as graphics is developed. The accuracy of proper text identification revealed a result of 91%
Keywords: Text, detection, optical character recognition, scene,detector
DOI: https://doi.org/10.5281/zenodo.7620902
PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT22DEC044_(1).pdf
REFERENCES