🚦 German Traffic Sign Classification

This project implements a Convolutional Neural Network (CNN) to classify images of German traffic signs . The model can help interpret traffic signs for autonomous driving systems.

---

📁 Dataset

We use the GTSRB - German Traffic Sign Recognition Benchmark, available on Kaggle:

• Train set: Images are stored in class-labeled folders (0 to 42).
• Test set: Images are in one folder, and labels are provided in a CSV file Test.csv.

Each image is resized to 64x64 and loaded using OpenCV.

---

🛠️ Features

• Load and preprocess images
• Resize all to a unified shape (64x64)
• CNN model building and training
• Early stopping and validation split
• Evaluation on test set
• Save predictions to CSV

---

🔧 Installation

pip install numpy pandas matplotlib opencv-python pillow scikit-learn tensorflow

Or simply run in Kaggle or Google Colab where most libraries are preinstalled.

---

📊 Results

• Final Accuracy: 96.3%
• Early stopping to avoid overfitting
• Confusion matrix and sample predictions visualized in the notebook

---

📈 Project Workflow

1. Load & Preprocess Data
2. Build CNN Model
3. Train the Model with Validation
4. Evaluate Accuracy and Loss
5. Visualize Results

---

🔍 Example Output

• Training vs. Validation Accuracy
• Training vs. Validation Loss
• Model performance on test set
• Sample predictions with true vs. predicted labels

---

👩‍💻 Author

Mariam Badr
Faculty of Computers & Artificial Intelligence, Cairo University
GitHub – LinkedIn