Introduced in MONet: Unsupervised Scene Decomposition and Representation2019
The Multi-Object Network (MONet) is developed, which is capable of learning to decompose and represent challenging 3D scenes into semantically meaningful components, such as objects and background elements.
