Deep convolutional inverse graphics network (DC-IGN) is a model that aims to learn an interpretable representation of images, disentangled with respect to three-dimensional scene structure and viewing transformations such as depth rotations and lighting variations. As the below image suggests, the deep convolutional inverse graphics network has a model that includes an encoder and a decoder -- it is a type of neural network that uses various layers to process input to output results. The DC-IGN example shows that it uses initial layers to encode through various convolutions, utilizing max pooling, and then uses subsequent layers to decode with unspooling.
