Publication

Abstract : Handwritten Text Recognition (HTR) holds significant importance in modern information processing, bridging the gap between the tangible world of handwritten content and the digital realm. With the rise of digitization, HTR plays a pivotal role in various domains, transforming handwritten text into machine-readable and editable formats. HTR serves as a transformative bridge that makes handwritten content accessible, editable, and useful in the digital age. There were many advancements in recognizing normal handwriting, but cursive-style writing especially for doctors is limited. In this paper, we have proposed an improved version of cursive HTR with a higher accuracy rate and fast processing using the CRNN Architecture. We have also added a spelling correction module which boosts the accuracy rate along with Multilingual support and text-to-speech conversion. We have achieved an accuracy of 96.51 percent with the help of the spelling correction module.