Exploration and Research on OCR Technology for Standardized Yi Character

Abstract

Optical Character Recognition (OCR) refers to the process of analyzing, understanding, and recognizing textual information in image files, which is beneficial for extracting and collecting standardized Yi text information from paper materials. Using Tesseract-OCR software, based on the coding standards of standardized Yi characters and the font characteristics of standardized Yi characters, a deep learning training and recognition result validation were conducted using Long Short Term Memory Neural Network (LSTM) for standardized Yi character recognition. The Python programming language called OpenCV library, Pytesseract library, and PyQt5 library to achieve the construction of a standardized Yi character recognition system. Experiments show that the system can realize the characters recognition of Yi characters in Baiti, Songti, Heiti and Xiheiti, and has off-line operation and high recognition accuracy.