Tai Le Character Recognition Using Deep Convolution Neural Networks

Abstract

Tai Le characters are used widely in Southwest China and Southeast Asia, and it is necessary to recognize Tai Le characters. In this paper, a Tai Le characters recognition model base on the deep convolution neural network (DCNN) has been established. In addition, to recognize Tai Le characters, the printed Tai Le database YDH2019.1 has been constructed, and image digital processing technology has been used to preprocess the sample images of YDH2019.1. In the modeling of DCNN, 50% data of YDH2019.1 has been used as train data, and the performance of the recognition, model has been evaluated on the remaining 50%. The results show that the recognition accuracy of Tai Le characters reaches 99.98%, which proved the effectiveness of DCNN as a printed Tai Le character recognition model.