Reweighted sparse representation with residual compensation for 3D human pose estimation from a single RGB image

Jiang, Mengxi; Yu, Zhuliang; Zhang, Yan; Wang, Qicong; Li, Cuihua; Lei, Yunqi

doi:10.1016/j.neucom.2019.05.034

Cited by 17 publications

(3 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the help of mature technology for estimating 2D keypoints, researchers estimate 3D poses on this basis, such as [4], [8], [9], [19]- [25]. Wang et al [19] notice the importance of the occlusion relationship of the joints, which is used as part of the loss for the network training.…”

Section: Related Workmentioning

confidence: 99%

“…Recent studies [4], [22]- [25] take into account the time information and use video as input instead of a separate frame. Pavllo et al [4] regard the process of estimating 3D position from 2D as the encoding part.…”

Section: Related Workmentioning

confidence: 99%

“…Then, the hidden information extracted from the picture is combined to directly obtain 3D poses. Jiang et al [25] pretrain a series of 3D poses using input pictures labeled with 2D joint information. They take the weights of these poses from the model to get a rough pose and use the residual compensation to obtain the final predicted pose.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Motion Projection Consistency Based 3D Human Pose Estimation with Virtual Bones from Monocular Videos

Wang¹,

Honghao²,

Wang³

et al. 2021

Preprint

View full text Add to dashboard Cite

Real-time 3D human pose estimation is crucial for human-computer interaction. It is cheap and practical to estimate 3D human pose only from monocular video. However, recent bone splicing based 3D human pose estimation method brings about the problem of cumulative error. In this paper, the concept of virtual bones is proposed to solve such a challenge. The virtual bones are imaginary bones between non-adjacent joints. They do not exist in reality, but they bring new loop constraints for the estimation of 3D human joints. The proposed network in this paper predicts real bones and virtual bones, simultaneously. The final length of real bones is constrained and learned by the loop constructed by the predicted real bones and virtual bones. Besides, the motion constraints of joints in consecutive frames are considered. The consistency between the 2D projected position displacement predicted by the network and the captured real 2D displacement by the camera is proposed as a new projection consistency loss for the learning of 3D human pose. The experiments on the Human3.6M dataset demonstrate the good performance of the proposed method. Ablation studies demonstrate the effectiveness of the proposed inter-frame projection consistency constraints and intra-frame loop constraints.

show abstract