Vivoice - Reading Assistant for the Blind using OCR and TTS

Vision is among the key senses that contribute greatly to the daily work. Visually disabled people rely on their other senses to do what they need. Lessened and missing vision can affect all facets of life, including ordinary personal activities, interaction with the community, prospects for school and work and the ability to acquire public services. Various designs were built from the 1970s to assist the blind. This paper describes our project which is designed to help the blind people to recognize various texts in their surroundings and also helps them to identify different objects around them. The device consists of a Raspberry Pi 3 module which acts as a System on Chip (SoC), a camera, push switch and earphones. The SoC is powered by a portable power bank. It has a CMOS camera, from which the images are captured and provided as an input for further processing. Optical Character Recognition is the techniques used for extracting the required texts from the images and it is further processed for obtaining the output. The images are processed with OpenCV coded in python using the Tesseract OCR library. The texts extracted are converted into a voice using Text-to-Speech synthesizer. eSpeak is the software used for text-to-speech conversion. The final output is given to the visually impaired people through the earphones. Another application employed is by the use of Natural Language Processing Algorithm for searching the necessary product by giving it as an input to the device. The device searches for the product and intimates it to the visually impaired person. By this device, the visually impaired people can reduce their dependence on the other people and their other senses and would be able to cater to their daily needs.