A Comprehensive Review for Optical Character Recognition of Handwritten Devanagari Script

Devanagari is a popular script in the Indian continent and is utilized for languages like Hindi, Marathi, Konkani, etc. A substantial part of the Indian population is unfamiliar to the English dialect and still uses Devanagari script to note the major documents even now. Recognition and digitization of this script will aid in various fields like banking and government departments. Thus, there is a demand to have a dependable application to recognize handwritten Devanagari script. Handwritten Character Recognition deals with identifying human written characters to convert them into digital text. Recognition of Devanagari Handwritten Characters is challenging in contrast to the Roman or English characters recognition because of the existence of a header line shirorekha. It is used to link the Devanagari characters in the formation of a word. Besides, the huge difference in the form of writing increases the intricacy. A lot of research has been done in this field; however, there is still a scope for enhancing the performance. This paper provides an extensive review of the various techniques used for the recognition of Devanagari Handwritten Character Recognition. It systematically analyzes various traditional machine learning techniques and deep learning approaches used in this domain. The paper also studies the advantages and disadvantages of each technique as well as discusses the challenges to be resolved for an effective and precise Devanagari Handwritten Character Recognition method.