Computer vision

Camera vision

Camera inputs

Classifying images

Semantic image segmentation

Building 3D models

Multi-view CNNs

Volumetric models

Point clouds

Polygon mesh

Generative Query Network (GQN)

Primitive-based CAD

3D ShapeNets

Building 3D models from scans



Classification with voxels

Absolute risk aversion

Correspondence (2 models of humans in different poses; understand where each part of one relates to a specific part of the other)


Mulitple objects in scene, objects have parts segmentation.