Sign Language Translator
Project Description
The main purpose of project is translating sign-language to letters. In accordance with this purpose we deice a dataset and we trained a CNN-Model with the dataset. You can find the summary of the model. We proceed the image that come from camera thanks to openCV and we predicted with CNN-Model. We take advantage of PyQT5 to provide a graphical user inetrafce to user.
Project Features
Sign-Language Translator
Education of Sign-Language
Versions
Current Project version : v0.8.0
Current model version : v0.3.0
Python version : v3.8.8
First of all you need to download python at least 3.8.8, you can download from Installation And Executionhere.
Then you can use the requirements.txt file to download necessary python libaries.
Then you will be have necessary python libaries, so you can run the program with that following command. pip install -r requirements.txt
python HomePage.py
How To ?
First of all you need to start project from Home.py. When you start the program a user graphical user interface pop up, In this interface there are two main and one assistan feature;
1-)Sign-Language Translator
The application translate the sign-language to letters that user show specified area. There are some features that established in this window as follows;
1.1-)Did you mean
After translation process is done, User can click "did you mean" button to list all of words that similar to predicted word. User can select any words that she/he wants and thanks to this feature we avoid wrong translations.
1.2-)Direct select to letter
While translating letters user can directly select a letter from list without waiting for prediction time.
2-)Education
The application also provide an education system to gain ability about Sign-Language. In this mode first of all user needs to select a letter that want to train, after selection an image appers this image represents as selected letter in terms of Sign-Language.
3-)Dataset
User can see the letters in terms of Sing-Language, This window does not close until user close thanks to that, user can access data while translating some letters.