High-performance Stacked denoising Autoencoder enhanced by CrossPre (Cross-Layer Pretraining). – K.Nakayama