On the paper-based information, a simple contextual and constant font style and size represent the presentation of contemporary lucrative OCR algorithms [1417]. To increase the accuracy, a huge number of datasets are created to train deep learning algorithms, neural networks (NN) as well as the