[CVPR'20 Oral] Self-Supervised 3D Human Pose Estimation via Part Guided Novel Image Synthesis (long)

ECCV 2020

Our framework disentangles factors of variation in human-centric images via a joint-anchored puppet model without accessing any paired supervision. Project page: https://sites.google.com/view/pgp-human Paper: https://openaccess.thecvf.com/content_CVPR_2020/papers/Kundu_Self-Supervised_3D_Human_Pose_Estimation_via_Part_Guided_Novel_Image_CVPR_2020_paper.pdf Authors: Jogendra Nath Kundu, IISc Bengaluru Siddharth Seth, IISc Bengaluru Varun Jampani, Google Research Mugalodi Rakesh, IISc Bengaluru R. Venkatesh Babu, IISc Bengaluru Anirban Chakraborty, IISc Bengaluru Abstract: Camera captured human pose is an outcome of several sources of variation. Performance of supervised 3D pose estimation approaches comes at the cost of dispensing with variations, such as shape and appearance, that may be useful for solving other related tasks. As a result, the learned model not only inculcates task-bias but also dataset-bias because of its strong reliance on the annotated samples, which also holds true for weakly-supervised models. Acknowledging this, we propose a self-supervised learning framework to disentangle such variations from unlabeled video frames. We leverage the prior knowledge on human skeleton and poses in the form of a single part based 2D puppet model, human pose articulation constraints, and a set of unpaired 3D poses. Our differentiable formalization, bridging the representation gap between the 3D pose and spatial part maps, not only facilitates discovery of interpretable pose disentanglement, but also allows us to operate on videos with diverse camera movements. Qualitative results on unseen in-the-wild datasets establish our superior generalization across multiple tasks beyond the primary tasks of 3D pose estimation and part segmentation. Furthermore, we demonstrate state-of-the-art weakly supervised 3D pose estimation performance on both Human3.6M and MPI-INF-3DHP datasets. Citation: @inproceedings{kundu2020self, title={Self-Supervised 3D Human Pose Estimation via Part Guided Novel Image Synthesis}, author={Kundu, Jogendra Nath and Seth, Siddharth and Jampani, Varun and Rakesh, Mugalodi and Babu, R Venkatesh and Chakraborty, Anirban}, booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition}, pages={6152--6162}, year={2020} }