Mixed Autoencoder for Self-Supervised Visual Representation Learning | IEEE Conference Publication | IEEE Xplore