Translation of Multilingual Text into Speech for Visually Impaired Person

Today’s human contact is primarily through voice and text. A person must have a proper vision to view and access the information available in the form of text. In any case, individuals who are blind can gather information by using their hearing ability. The proposed method is a camera-based assistive message reading system that allows blind people and travellers to read the E-messages, printed notes, and other materials in their native languages. It integrates the ideas of optical character recognition (OCR), text to voice synthesizer (TTS), and interpreter in the Raspberry Pi. The proposed model coverts the text from verified records or a caption overlay available on an image into machine-readable text. Text-to-Speech conversion has been enabled by employing the OCR operation to sweep and read any language characters and numbers present in an image and interpret the information in any desired language, and lastly produce a voice output of the decoded text. Voice is the output obtained through the Raspberry Pi’s sound jack, which may be heard by using speakers or headphones. The main goal of this research work is to develop and implement a novel model to transcribe the Text to Speech in many dialects.

I. Introduction

In recent years, machine reading has advanced from a research vision to reality. People who are visually challenged experience a wide range of difficulties in accessing the written text with modern technology, including the issues with layout, exactness, flexibility, and productivity. This research study presents a smart technique to assist both visually challenged people and travellers by reading paper-printed messages successfully. The proposed structure is built on deploying image acquiring technologies in an embedded framework on the Raspberry Pi board. The concept is based on preliminary research with people who are visibly impaired, and it is small and portable, allowing for a more appropriate activity with less planning. For voyagers and the visibly weaker, we presented a text-read-out framework in this project. A camera serves as an information device in the proposed fully integrated framework, which is used to manage the printed text record for digitalization. Discourse is perhaps the most effective way for the individuals to communicate with each other. An optical character recognition approach is used to remove the text from the image. Optical character recognition (OCR) is a process that translates verified or printed text, as well as manually typed data, into editable text for further processing. Discourse mixture is a phoney mix of human conversation. A Text-to-Speech (TTS) synthesiser is a computer framework that can read any text and make it audible to everyone, regardless of whether it was simply input into the computer by an admin or reviewed and sent to an Optical Character Recognition (OCR) framework. The device was put through its paces on the Raspberry Pi platform. The Raspberry Pi is a basic installed framework that, as a low-cost single-load up PC, is used to reduce the complexity of frameworks as applications grow. Python is the foundation of this level. The text is physically centred on the Pi camera. Then it takes a picture, with a 5 second delay to help centre the pi camera if it is accidentally defocused. After a delay, Raspy takes the image and processes it such that you may hear the message’s deciphered expressions over the headphones or speaker linked to Raspy through its 3.5mm sound connector.