Geometry-Supervised-Pose-Network-for-Accurate-Retail-Shelf-Pose-Estimation

Monocular shelf posture estimation algorithm in the retail field

Overview

Code release for Geometry Supervised Pose Network for Accurate Retail Shelf Pose Estimation. This is a GSPN which contains a GSL Module. Using VGG as backbone, I add GSL module into vgg as a monitor to supervise the shelf pose estimatiom, a complete pose of shelf containing three value in 3D space.

This paper is jointly complished by Yongqiang Mou *，Zhiyi Huang (JXSTNUAA / AIBC)，Lingfan Lin ， Yishi Guo and Zhen Yang (JXSTNUAA / AIBC) in IEEE Transactions on Industrial Informatics (IEEE TII, IF=9.112) 2020。

Further information please contact Yongqiang Mou

Requirements

PyTorch (version >= 1.1.0) Torchvision (version >= 0.3.0)

Getting Started (Testing)

Testing

Model-Site:

链接:https://pan.baidu.com/s/1Pmi6bfbDok0VOlKjMOGDLA  
密码:uhnz

Command:

cd ~/project  
mkdir models
cp ./checkpoint.pth.tar ~/project/models
download datasets
mv ./traindata ~/project/dataset
cd ~/project/code
python python eval.py

Experimental Result

Basis	Method	Input	Pitch Model	Yaw Model	Roll Model	Average Error
FFPVA[1]	FFPVA[1]	RGB	16.497◦	6.574◦	5.087◦	9.386◦
PoseNet[2]	BL-RGB	RGB	8.533◦	2.991◦	1.345◦	4.289◦
PoseNet[2]	BL-LSD	GIM	8.083◦	2.662◦	1.125◦	3.957◦
PoseNet[2]	BL-Fusion	RGB+GIM	8.078◦	2.609◦	1.102◦	3.929◦
L2SD	L2SD	RGB	7.958◦	1.543◦	1.119◦	3.540◦
GSPN	GSPN	RGB	7.456◦	0.864◦	0.761◦	3.027◦

Reference

[1]. J. Lezama, and G. Randall R. G. V. Gioi, and J. M. Morel, “Finding vanishing points via point alignments in image primal and dual do�mains,” IEEE Computer Vision and Pattern Recognition, pp. 509–515, Jun. 2014.
[2]. A. Kendall, M. Grimes, and R. Cipolla, “Posenet: A convolutional network for real-time 6-dof camera relocalization,” IEEE International Conference on Computer Vision, p. 265, Dec. 2015.

License and Citation

@article{0Geometry, title={Geometry Supervised Pose Network for Accurate Retail Shelf Pose Estimation}, author={Mou, Yongqiang and Huang, Zhiyi and Lin, Lingfan and Guo, Yishi and Yang, Zhen}, journal={IEEE Transactions on Industrial Informatics}, volume={PP}, number={99}, pages={1-1}, year={2020} }

Huang9495/GSPN