DensePose: Dense Human Pose Estimation In The Wild

Abstract : In this work, we establish dense correspondences between an RGB image and a surface-based representation of the human body, a task we refer to as dense human pose estimation. We gather dense correspondences for 50K persons appearing in the COCO dataset by introducing an efficient annotation pipeline. We then use our dataset to train CNN-based systems that deliver dense correspondence ‘in the wild’, namely in the presence of background, occlusions and scale variations. We improve our training set’s effectiveness by training an inpainting network that can fill in missing ground truth values and report improvements with respect to the best results that would be achievable in the past. We experiment with fully-convolutional networks and region-based models and observe a superiority of the latter. We further improve accuracy through cascading, obtaininga system that delivers highly-accurate results at multiple frames per second on a single GPU. Supplementary materials, data, code, and videos are provided on the project page
Liste complète des métadonnées
Contributor : Riza Alp Guler <>
Submitted on : Tuesday, December 11, 2018 - 4:49:24 PM
Last modification on : Monday, March 4, 2019 - 2:44:13 PM
Document(s) archivé(s) le : Tuesday, March 12, 2019 - 3:32:14 PM


Files produced by the author(s)


  • HAL Id : hal-01951864, version 1


Riza Alp Güler, Natalia Neverova, Iasonas Kokkinos. DensePose: Dense Human Pose Estimation In The Wild. Conference on Computer Vision and Pattern Recognition (CVPR) 2018, Jun 2018, Salt Lake City, United States. ⟨hal-01951864⟩



Record views


Files downloads