Devanagari Character Recognition Using Convolutional Neural Network

  • Priyanka Deshpande, Ankita Bondre, Vaishnavi Bhise, H.B.Ghorpade ,S.S.Bhosale, C.G.Patil


In the last few years, we can see many articles have published on optical character recognition (OCR). Nowadays there are many commercial OCR systems present in the market, most of these systems work for Roman, Chinese, Japanese, and Arabic characters. In India, we have 12 different scripts but there is no sufficient work is done on Indian language character recognition. In this paper, we are focusing on the Devanagari script which is used in different languages like Hindi, Sanskrit, Nepali, Marathi, Sindhi, and Konkani. Recently deep learning is replacing other pattern recognition techniques. Deep learning is the right choice to handle the challenges are involved in applications like character recognition, which involve a large amount of the database and variability in the data, we are using the Convolutional Neural Network for the feature extraction and the classification followed by pre-processing. This approach has been proven effective and accurate in identifying characters in a digital format.