Until now, we have seen that autoencoder inputs are images. So, it makes sense to ask whether a convolutional architecture can work better than the autoencoder architectures discussed previously.

We will analyze how the encoder and decoder work in convolutional autoencoders.