A Comprehensive Study of Weight Sharing in Graph Networks for 3D Human Pose Estimation

Liu, Kenkun; Ding, Rongqi; Zou, Zhiming; Wang, Le; Tang, Wei

doi:10.1007/978-3-030-58607-2_19

Cited by 105 publications

(54 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Martinez et al [23] propose a simple yet effective baseline for 3D human pose estimation, it uses only 2D joints as input but gets highly accurate results, showing the importance of 2D joints information for 3D human pose estimation. Since the skeleton's topology can be viewed as a graph structure, there has been increasing use of Graph Convolutional Networks (GCNs) for 2D-to-3D pose estimation tasks [4,19,21,37,40,41].…”

Section: Related Workmentioning

confidence: 99%

“…where is an element-wise product operation and ρ i (•) is softmax nonlinearity which normalizes the input matrix across all choices of i. Following previous works [21,38,41], we use two different transformation matrices for the representation of each node i and its neighbors respectively in actual implementation.…”

Section: Preliminariesmentioning

confidence: 99%

“…The sparsest graph for the bottom subnetwork in Figure 2 is defined on the standard human skeletal graph G P = {V P , E P }. Similar to existing methods [21,23,44], the number of nodes J is set to 17 in this paper. For the middle and top subnetworks, we coarsen the human mesh graph M whose the number of nodes V c equal to 96 and 48, respectively, to maintain a suitable model size.…”

Section: Hierarchical Graph Networkmentioning

confidence: 99%

“…Recent works show that using such detected 2D joints positions, the 3D human pose can also be efficiently and accurately regressed [6,10,23,29,36]. To further boost performance, many attempts have been made to explicitly utilize the human skeletal topology and use graph convolutional networks (GCNs) to exploit the spatial configurations for 3D human pose estimation [4,21,37,41]. However, the graph topology of human skeleton is usually formed by few number of joints (e.g.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Hierarchical Graph Networks for 3D Human Pose Estimation

Shi

Dai

et al. 2021

Preprint

View full text Add to dashboard Cite

Recent 2D-to-3D human pose estimation works tend to utilize the graph structure formed by the topology of the human skeleton. However, we argue that this skeletal topology is too sparse to reflect the body structure and suffer from serious 2D-to-3D ambiguity problem. To overcome these weaknesses, we propose a novel graph convolution network architecture, Hierarchical Graph Networks (HGN). It is based on denser graph topology generated by our multi-scale graph structure building strategy, thus providing more delicate geometric information. The proposed architecture contains three sparse-to-fine representation subnetworks organized in parallel, in which multi-scale graph-structured features are processed and exchange information through a novel feature fusion strategy, leading to rich hierarchical representations. We also introduce a 3D coarse mesh constraint to further boost detail-related feature learning. Extensive experiments demonstrate that our HGN achieves the state-of-the-art performance with reduced network parameters.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Preliminariesmentioning

confidence: 99%

Section: Hierarchical Graph Networkmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Hierarchical Graph Networks for 3D Human Pose Estimation

Shi

Dai

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Ci et al [27] present a locally connected network to improve the representation capability of GCNs. Liu et al [28] first systematically analyze the mechanism of weight sharing in graphs. Inspirited by them, we apply GCNs to 3D human pose prediction.…”

Section: B Graph Convolutional Networkmentioning

confidence: 99%

3D Human Pose Estimation With Spatial Structure Information

Huang

Tang

2021

IEEE Access

View full text Add to dashboard Cite

Estimating 3D human poses from 2D poses is a challenging problem due to joints selfocclusion, weak generalization, and inherent ambiguity of recovering depth. Actually, there exists spatial structure dependence on human body key points which can be used to alleviate the problem of joints selfocclusion. Therefore, we represent human pose as a directed graph and propose a network implemented with graph convolution to predict 3D poses from the given 2D poses. In the digraph, we determine the connection weight of each edge according to the error distribution of joints estimation. This makes our model robust to noise. By optimizing coarse 3D estimation and adversarial learning, our algorithm can successfully improve the accuracy of estimation and relieve the ambiguity of mapping. Through testing on Human 3.6M and MPI-INF-3DHP datasets, we achieve excellent quantitative performance. More importantly, our algorithm also has a superior generalization to outdoor dataset MPII by the pre-training process.INDEX TERMS 3D human poses, graph convolutional networks, adversarial learning, geometric priors, gradient vanish, in-the-wild scenes.

show abstract