Details for entry PGPT

Short Name:
PGPT
Long Name:
Pose-Guided Tracking-by-Detection: Robust Multi-Person Pose Tracking
Use MPII Pose for training:
no
Use COCO for training:
yes
Use external training:
no
Use multiple frames for pose estimation:
yes
Predict pose tracks:
yes
URL:
https://ieeexplore.ieee.org/document/9034193
Description:
We propose a pose-guided tracking-by-detection framework that fuses pose information into both video human detection and human association procedures. In the video human detection stage, we adopt the pose-guided person location prediction exploiting the temporal information to make up missing detections. Technically, pose heatmaps are utilized to cope with the person-specific intra-class distractors. Furthermore, in the human association stage, we propose an appearance discriminative model based on the hierarchical poseguided graph convolutional networks (PoseGCN). The PoseGCNbased model exploits the human structural relations to boost the person representation.
Bibtex:
@ARTICLE{PGPT, author={Q. {Bao} and W. {Liu} and Y. {Cheng} and B. {Zhou} and T. {Mei}}, journal={IEEE Transactions on Multimedia}, title={Pose-Guided Tracking-by-Detection: Robust Multi-Person Pose Tracking}, year={2020}, }
Specs:
20GHz
Runtime:
2400
Open Source:
yes
Number of submissions:
4
Last Submission:
2019-03-21 05:13:29
Published:
2018-11-26 08:48:30

Imprint | Data Protection


© PoseTrack