Self-Supervised Visual Representation Learning from Hierarchical Grouping

Xiao Zhang, Michael Maire

We create a framework for bootstrapping visual representation learning from a primitive visual grouping capability. We operationalize grouping via a contour detector that partition…