Skip to main content
. 2019 Jul 1;26(11):1286–1296. doi: 10.1093/jamia/ocz089

Figure 4.

Figure 4.

The original images (blue rectangle) and the corresponding reconstructed images (red rectangle) by autoencoders in 3 different models. Model 1: an autoencoder using labeled images. Model 3: an autoencoder + a classifier (Clf) using only labeled images. Model 4 (Convolutional AutoEncoder based Semi-supervised Network): an autoencoder + a classifier using both labeled and unlabeled images. Models 1 and 4 (Convolutional AutoEncoder based Semi-supervised Network) achieve better reconstruction results compared with those of model 3.