Turkish Sign Language Characters Recognition

CNN Model training with desktop GUI for recognition turkish sign language characters and using trained model with mobile app.

Used Tools

• TensorFlow, TensorFlowLite, TensorBoard
• PyQt5
• NumPy
• OpenCV
• Matplotlib
• seaborn

Run Project

python isaredilialgila.py

Features

• Dataset of approximately 4500 images.
• GUI for model training and testing.
• Custom model training.
• Transfer learning with MobileNetV2, ResNet50, InceptionV3.
• Model using with android app.
• Multiple model storeing.
• Model visualization with TensorBoard.
• Real time model testing with OpenCV.
• Model transformation for use with TensorFlowLite.

⚠️ For more details: Report

Screenshots

Desktop App

Mobile App (Android)

License

MIT