This Keras code builds and trains a CNN for MNIST digit classification. The model includes three convolutional layers, each followed by max-pooling, and two dense layers. It achieves 99.00% accuracy on the test set. The dataset is normalized and labels are one-hot encoded.