Enhancing Email Accessibility for the Visually Impaired through Deep Neural Network

The Interactive Voice Response (IVR) application based on Deep Neural Networks (DNN) and Recurrent Neural Networks (RNN) allows persons with vision impairments or socially less privileged, uneducated, or handicapped users quickly and effectively utilize the operating system's multimedia and email features. Studies show that over 230 million people across the world are unaware of the Internet or email. A large group of them are either blind or illiterate. Email platforms often feature intricate user interfaces with numerous buttons, menus, and settings, which is challenging. This paper presents the voice mail architecture that permits blind, uneducated, or handicapped users to read and send emails using a voice- based mailing service deprived of assistance from their G-mail accounts. The usage of the keyboard is eliminated, and only mouse and voice input are utilized to perform E-mail options. Advanced Text to speech framework using a vocoder module and an acoustic model, based on DNN and RNN, is employed to read the email text with enhanced accuracy, speed, and low Word Error Rate (WER).