Bottom-Up Human Pose Estimation Via Disentangled Keypoint Regression
Authors:
Top Cited Papers
- 1 June 2021
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 14671-14681
Abstract
In this paper, we are interested in the bottom-up paradigm of estimating human poses from an image. We study the dense keypoint regression framework that is previously inferior to the keypoint detection and grouping framework. Our motivation is that regressing keypoint positions accurately needs to learn representations that focus on the keypoint regions.We present a simple yet effective approach, named disentangled keypoint regression (DEKR). We adopt adaptive convolutions through pixel-wise spatial transformer to activate the pixels in the keypoint regions and accordingly learn representations from them. We use a multi-branch structure for separate regression: each branch learns a representation with dedicated adaptive convolutions and regresses one keypoint. The resulting disentangled representations are able to attend to the keypoint regions, respectively, and thus the keypoint regression is spatially more accurate. We empirically show that the proposed direct regression method outperforms keypoint detection and grouping methods and achieves superior bottom-up pose estimation results on two benchmark datasets, COCO and CrowdPose. The code and models are available at https://github.com/HRNet/DEKR.Keywords
All Related Versions
This publication has 57 references indexed in Scilit:
- Human Pose Estimation with Iterative Error Feedback2016
- Multi-person Pose Estimation with Local Joint-to-Person Associations2016
- DeeperCut: A Deeper, Stronger, and Faster Multi-person Pose Estimation Model2016
- Human Pose Estimation Using Deep Consensus Voting2016
- Human Pose Estimation via Convolutional Part Heatmap Regression2016
- Robust Optimization for Deep Regression2015
- DeepPose: Human Pose Estimation via Deep Neural Networks2014
- Poselet Conditioned Pictorial Structures2013
- Representation Learning: A Review and New PerspectivesIEEE Transactions on Pattern Analysis and Machine Intelligence, 2013
- Articulated pose estimation with flexible mixtures-of-parts2011