Project description

This is the backend for a speech-to-signLanguage app making use of Natural Language Processing and the Google speech-to-text API. The final translation is shown as a video with subtitles. If a word is not available in the database then the word is displayed on a black screen.

The frontend is written in React.

Getting started

The following guide explains how to set up this project.

Set up the database

Prerequisites

• Install ImageMagick as it is needed to write texts with the moviepy library.
• Clone WLASL Git repository

Note: WLASL Git repository is used to retrieve the open-source videos for the database.

Download database content

To get the database you must configure your parameters first. The first needed parameter is your local file path to WLASL_v0.3.json from WLASL Git repository. The second parameter is the file path to the required database destination. Finally, run the DatabaseBuilder.py in the start_kit folder under code.

Run project

Prerequisites

• Install the database as described above.
• Place your googleCloud key under codeBase/keys.

Final step

Run the frontend and main.py under codeBase.

Run tests

To run test, please add your own audio files.