Machine and Deep Learning Classifiers for Indic Scripts Recognition: Challenges and Research Perspective

The character recognition technique is based on electronic or mechanical transformation of handwritten or printed document images into machine interpretable form. The machine learning and deep learning based methods have been proven to be highly effective for the character recognition problems. In this work, the main objective is to review and explore different classification methods available for the recognition of handwritten Bangla and Devanagari scripts. The classification models are broadly categorized into three groups, viz., SVM, neural networks, and transfer learning based techniques.