Rotation-invariant Mixed Graphical Model Network for 2D Hand Pose Estimation

Deying Kong, Haoyu Ma, Yifei Chen, Xiaohui Xie; The IEEE Winter Conference on Applications of Computer Vision (WACV), 2020, pp. 1546-1555

Abstract


In this paper, we propose a new architecture named Rotation-invariant Mixed Graphical Model Network (R-MGMN) to solve the problem of 2D hand pose estimation from a monocular RGB image. By integrating a rotation net, the R-MGMN is invariant to rotations of the hand in the image. It also has a pool of graphical models, from which a combination of graphical models could be selected, conditioning on the input image. Belief propagation is performed on each graphical model separately, generating a set of marginal distributions, which are taken as the confidence maps of hand keypoint positions. Final confidence maps are obtained by aggregating these confidence maps together. We evaluate the R-MGMN on two public hand pose datasets. Experiment results show our model outperforms the state-of-the-art algorithm which is widely used in 2D hand pose estimation by a noticeable margin.

Related Material


[pdf]
[bibtex]
@InProceedings{Kong_2020_WACV,
author = {Kong, Deying and Ma, Haoyu and Chen, Yifei and Xie, Xiaohui},
title = {Rotation-invariant Mixed Graphical Model Network for 2D Hand Pose Estimation},
booktitle = {The IEEE Winter Conference on Applications of Computer Vision (WACV)},
month = {March},
year = {2020}
}