CodeAlpha_Handwritten-character-recognition

EMNIST + MNIST CNN Classifier

This project combines the MNIST (digits) and EMNIST Letters datasets to train a CNN model that classifies handwritten digits (0–9) and uppercase letters (A–Z) using TensorFlow and Keras.

---

📦 Dataset

• MNIST: Loaded using tensorflow.keras.datasets.mnist.
• EMNIST Letters: Downloaded from Kaggle using kagglehub. You must have your Kaggle API key configured.

---

🔧 Setup Instructions

1. Install Dependencies

```bash pip install numpy pandas matplotlib seaborn tensorflow scikit-learn kagglehub

Ensure you have your kaggle.json API key file placed correctly: mkdir ~/.kaggle cp kaggle.json ~/.kaggle/ chmod 600 ~/.kaggle/kaggle.json

Run the Script python train_model.py

Label Mapping Digits: 0–9 Letters: 10–35 → A–Z Total of 36 classes.

Evaluation Training/Validation Accuracy is plotted. Final test performance is printed, along with a classification report. Sample predictions are shown for visual inspection.

Example Output Test Accuracy: 0.9745 Sample Predictions: True: 0, Predicted: 0 True: A, Predicted: A True: 9, Predicted: 9

File Structure . ├── train_model.py # Main script ├── best_model.keras # Saved best model checkpoint └── README.md # This file