An Ancient Indian Handwritten Script Character Recognition by Using Deep Learning Algorithm

Abstract

Handwritten character recognition is one of the most popular and widely used techniques in Pattern Recognition and Image processing. The Difficulty level about the handwritten character recognition is increasing when the handwritten characters are the ancient script of any region. This offline handwritten character manual identification and analysis is a time consuming and lengthy process. During manual recognition process, there might have possibilities of human errors. This human error might change the meaning of existing ancient document content. Hence accurate analysis and transliteration/translation is an essential requirement for keeping ancient document alive. In this research paper, the ancient script called MODI script has used for handwritten characters identification. The overall recognition mechanism has presenting by a Classification algorithm which is a highly accurate Convolutional Neural Network (CNN) algorithm. The current paper is categorized into four sections. The First section describes the introductory part of the current research work along with Ancient MODI script and Deep learning model called AlexNet. The second section which is proposed system work like preparation of dataset, preprocessing, and data augmentation. The Next section is the Experimental arrangement and result of the proposed research work.The last section is the conclusive remark of proposed system.