Learning effective human pose estimation from inaccurate annotation.

Sam Johnson, Mark Everingham

Browse the full CVPR paper archive.