Design of Optimal Deep Learning Assisted Online Telugu Character Recognition Model

Abstract

Telugu character recognition process has received significant attention due to the exponential utilization of resources like images, smartphones, iPods, and paper documents. It can be divided into two types namely offline character recognition and online character recognition. Offline character recognition is a process of identifying Telugu characters from the scanned image or document whereas online character recognition enables to recognition of characters by the machine while the user writes. Several researchers have attempted to design online Telugu character recognition model by the use of distinct classification models and feature extraction approaches; however, the performance is yet to be improved. In this aspect, this study focuses on the design of optimal deep learning based online Telugu character recognition (ODL-OTCR) model. The goal of the ODL-OTCR technique is to recognize as well as classify the Telugu characters in online model. The ODL-OTCR technique involves data preprocessing to preprocess the character stroke in three ways namely normalization, smoothing, and interpolation. Besides, a beetle swarm optimization (BSO) with EfficientNet model is utilized as a feature extractor and finally, Siamese Neural Networks (SNN) model is employed for the classification process. In order to showcase the improved performance of the ODL-OTCR technique, a series of simulations take place and the results are inspected interms of different aspects. The simulation results highlighted the betterment of the proposed ODL-TCR technique over the recent techniques.