Unlock the power of communication with our Sign Language Detection project! This innovative project uses OpenCV and the MediaPipe Hand module to recognize American Sign Language (ASL) letters. With a dataset containing 100 samples for each of the 26 letters, the project is ready to go for ASL recognition. Whether you're directly detecting ASL or want to customize it for another alphabet, we've got you covered!
- opencv-python==4.7.0.68
- mediapipe==0.9.0.1
- scikit-learn==1.2.0
Install the requirements using:
pip install opencv-python mediapipe numpy tensorflow
To start recognizing ASL letters, simply run the camera_demo.py
file:
python camera_demo.py
If you want to train the model for a different alphabet, follow these steps:
-
Collect data by running
collect_imgs.py
:python collect_imgs.py
-
Process the collected images into a dataset using
create_dataset.py
:python create_dataset.py
-
Train your model with the new dataset using
train_model.py
:python train_model.py
-
Run the demo with the trained model using
camera_demo.py
. Ensure the model path incamera_demo.py
is correctly set.
Before running camera_demo.py
, make sure the model path is correctly set to the trained model in the script.
We welcome contributions! If you want to contribute, please send a pull request or open an issue. Your feedback and contributions are highly valued.