Sparse Autoencoder + dropout ?

I have solved UFLDL "Exercise: Sparse Autoencoders", but have trouble adding G. Hinton's Dropout technique + the maxout activation function. Somebody who can help me?

Task: http://deeplearning.stanford.edu/wiki/index.php/Exercise:Sparse_Autoencoder

Solution: sparse autoencoder (without Dropout or relu) example: https://github.com/onlymag4u/UFLDL_SparseAutoencoder/tree/master/sparseae_exercise/code