Deep Learning Approaches for Devanagari Handwriting Recognition
Shilpa Kaur Manocha, Piyush Tewari · 2021
Offline Handwriting Recognition, allows to recognize and convert physical files and images to a digital form. Handwriting recognition is a vast area of research and a lot of contribution has been witnessed in past decades. Devanagari Script is one of the scripts for which many sophisticated handwriting recognition systems are not available. Deep Learning techniques, however, have been able to generate excellent results and enhanced performance. This paper thoroughly examines the recent advancements in deep learning architectures to recognize Devanagari Handwritten characters from varying datasets. The paper has been focused on Devanagari Handwritten Recognition Systems background, challenges, opportunities and future research trends. Several publicly available datasets have also been mentioned along with their data size, type of characters. A comparative study has been done to present the results and insights of the state-of-the-art approaches.