Dynamic Keypoint Head

Computer VisionIntroduced 20001 papers

Description

Dynamic Keypoint Head is an output head for pose estimation that are conditioned on each instance (person), and can encode the instance concept in the dynamically-generated weights of their filters. They are used in the FCPose architecture.

The Figure shows the core idea. FF denotes a level of feature maps. "Rel. Coord." means the relative coordinates, denoting the relative offsets from the locations of FF to the location where the filters are generated. Refer to the text for details. f_θ_if\_{\theta\_{i}} is the dynamically-generated keypoint head for the ii-th person instance. Note that each person instance has its own keypoint head.

Papers Using This Method