Voice4mutes

TABLE OF CONTENTS

1. Overview

2. WebApp

   a) Login Page

   b) Learn and Translate

3. About ML/DL models

Overview

Our project is inspired by the current scenario/status of speech and hearing impaired people and how the biggest challenges faced by them are lack of social interaction, language and communication problems. We plan to reduce this divide. Voice4mutes powers individuals to overcome the barrier of communication using technology.

WebApp

a) Login Page

b) Learn And Translate

ML/DL Model

For our deeplearning models we have used ResNetV2 for transferlearning. We used openCV to make predictions and have made use of flask to deploy our model. The training accuracy of the model was 93% and the validation/testing accuracy is 88%. The training and testing split is 70:30.The model is best run in white background.