✍️ This project is part of my CodeAlpha Machine Learning Internship, where I built a system to recognize handwritten characters using Convolutional Neural Networks (CNNs) and the EMNIST dataset.
- Recognized handwritten alphabets with CNNs
- Trained and evaluated on EMNIST dataset
- Visualized predictions and model performance
- Achieved high accuracy in character classification
- MNIST Handwritten Digits dataset
- 60,000 training images and 10,000 testing images
- Images are grayscale, 28×28 pixels.
EMNIST Dataset — downloaded automatically via code during training from tensorflow.keras.datasets import mnist (X_train, y_train), (X_test, y_test) = mnist.load_data()
| Layer (type) | Output Shape | Parameters |
|---|---|---|
| Conv2D | (26, 26, 32) | 320 |
| MaxPooling2D | (13, 13, 32) | 0 |
| Flatten | (5408) | 0 |
| Dense (128) | (128) | 692,352 |
| Dense (10) | (10) | 1,290 |
Total Parameters: 693,962
- Python
- TensorFlow / Keras
- NumPy, Matplotlib, Seaborn
📈 Results Test Accuracy: ~99%
Sample Predictions:
Project submitted for CodeAlpha Machine Learning Internship