Akshat Santhana Gopalan
Created November 30, 2023

BeyondWords

AI Reading Companion for Individuals with Dyselxia

11
BeyondWords

Things used in this project

Software apps and online services

Jupyter Notebook
Jupyter Notebook
Google ML Kit
Google Text-To-Speech
OpenCV
OpenCV
React Native
Hugging Face
BART-CNN Model
Android Studio
Android Studio

Story

Read more

Schematics

Application Workflow Diagram

The app itself is built using the React Native Framework. After taking a picture of the text to read, the image is loaded with the help of the OpenCV Library. This goes to the Google Machine Learning Kit’s Optical Character Recognition Library, which extracts solely the text content from the image. Immediately, this text is converted to the OpenDyslexic text format, created to improve readability for people with dyslexia The app provides the ability to flexibly adjust text size & spacing too. The Google AI Text to speech library is also used with this extracted text to implement the read-aloud feature that the user can follow along. Intricate text is also simplified to make it shorter and more easily comprehensible using Hugging Face’s AI BART CNN Model, which combine a Bidirectional and Auto regressive transformer with a convolutional neural network to provide an efficient summary of the text.

Code

OpenCV BeyondWords

Credits

Akshat Santhana Gopalan

Akshat Santhana Gopalan

1 project • 1 follower

Comments