Batch Normalization - Is batch normalization better than dropout?

Batch normalization stabilizes the training process by lowering the internal covariate shift, while dropout assists in preventing overfitting by decreasing co-adaptation. Deep learning models can benefit from a combination of the two approaches.