Handwritten Character Recognition for Tamil Language Using Convolutional Recurrent Neural Network

Abstract

AbstractThe process of conversion of printed Tamil character from a handwritten Tamil character is known as handwritten Tamil character recognition. Handwritten characters are challenging to analyze because of the wide range of writing styles, as well as the varying sizes and orientation angles of the characters. This paper has the ability to recognize offline handwritten characters that presently supports Tamil language that has high accuracy and minimum validation loss by the system. The optical character recognition system developed for the Tamil language carry a poor recognition rate because of its numerous writing designs and enormous number of characters. The goal of this system is to notice a digital format for Tamil written document. The proposed method is to achieve a accuracy of ninety six percentage with the help of CRNN. It is the foremost difficult task even although, numerous researches have been planned for written Tamil text. The trendy analysis during this domain has used the deep learning algorithm rule to improve the accuracy for the input document. This survey grants associate in fair review of the Tamil written character recognition, challenges, and diverse skills.KeywordsHandwritten character recognitionWritten character recognitionTamil languageDeep learning techniquesConvolution recurrent neural networkHybrid neural networkArtificial neural network or deep neural network