English character recognition algorithm by improving the weights of MLP neural network with dragonfly algorithm

Abstract

Character Recognition (CR) is taken into consideration for years. Meanwhile, the neural network plays an important role in recognizing handwritten characters. Many character identification reports have been publishing in English, but still the minimum training timing and high accuracy of handwriting English symbols and characters by utilizing a method of neural networks are represents as open problems. Therefore, creating a character recognition system manually and automatically is very important. In this research, an attempt has been done to incubate an automatic symbols and character system for recognition for English with minimum training and a very high recognition accuracy and classification timing. In the proposed idea for improving the weights of the MLP neural network method in the process of teaching and learning character recognition, the dragonfly optimization algorithm has been used. The innovation of the proposed detection system is that with a combination of dragonfly optimization technique and MLP neural networks, the precisions of the system are recovered, and the computing time is minimized. The approach which was used in this study to identify English characters has high accuracy and minimum training time.