Build CNN: Apparel Image Classification for Fashion MNIST

This MATLAB script trains a CNN for MNIST digit classification, achieving >90% accuracy.

This MATLAB script trains a Convolutional Neural Network (CNN) on the MNIST dataset for digit classification. It follows a structured workflow:

1. Data Extraction & Processing – Decompresses and loads MNIST images and labels.
2. Visualization – Displays sample images from training and test sets.
3. CNN Architecture – Uses convolutional, batch normalization, ReLU, and pooling layers.
4. Training – Optimized with Adam optimizer and early stopping.
5. Evaluation – Computes accuracy, confusion matrix, precision, recall, and F1-score.