This project implements a Convolutional Neural Network (CNN) using TensorFlow/Keras to classify handwritten digits from the MNIST dataset. The model achieves 98.68% accuracy by utilizing data augmentation and optimization techniques.
- Dataset: MNIST (60,000 training images, 10,000 test images)
- Model Architecture: CNN with convolutional, pooling, and fully connected layers
- Data Augmentation: Image rotation to improve generalization
- Optimization: Batch normalization, dropout, and Adam optimizer
- Evaluation: Confusion matrix, accuracy, and loss metrics