This project is a Hand Gesture Recognition System built using a CNN model to classify hand gestures in real-time.
The system is designed for accuracy, stability, and smooth performance.
- Real-time hand gesture detection.
- CNN model optimized for stability and inference speed.
- Supports multiple gesture classes.
- Clean and modular project structure.
We chose a Convolutional Neural Network (CNN) because:
- It captures local spatial features in hand images effectively.
- It offers the best balance between accuracy and inference speed.
- More stable during training compared to other models.
- Works smoothly in real-time applications.
Here are the final metrics of the model:
| Metric | Value |
|---|---|
| Loss | 0.0737 |
| Accuracy | 0.9957 |