TAMIL-BRAHMI SCRIPT CHARACTER RECOGNITION SYSTEM USING DEEP LEARNING TECHNIQUE

Character recognition is the process of detecting and processing a character or words and storing it in textual format. This project aims to build a character recognition system to digitize the ancient Tamil script and additionally translate the obtained script to modern Tamil. Although technologies around image processing are almost at perfection for languages like English and other modern languages, recognition of ancient scripts is still elusive. Conventional methods used by the Archaeological department convert the documents manually, which is possible only with the help of an expert. The proposed system will cross out such a need and will enable the common man to decipher the ancient scripts from inscriptions and manuscripts and also digitize, and also enable to process the data using Natural Language Processing. The system incorporates Convoluted Neural Networks to extract the features from each character to detect and translate the character to modern Tamil with 94.6 % accuracy.