DGECN

This repo provides the PyTorch implementation of the work:

Tuo Cao, Fei Luo, Yanping Fu, Wenxiao Zhang, Shengjie Zheng, Chunxia Xiao. DGECN: A Depth-Guided Edge Convolutional Network for End-to-End 6D Pose Estimation. In CVPR 2022. Project page

Please note that this repo only contains test code without KFA and DRN, and test with RANSAC/PnP. But we also provide our DG-PnP, we may provide the full code in Huawei Ascend after the conference.

Overview

Requirements

• Ubuntu 16.04/18.04, CUDA 10.1/10.2, python >= 3.6, PyTorch >= 1.6, torchvision
• Generate your test file list as ycb-video-testlist.txt

Datasets and Pretrained model

Download the YCB-V from here and extract to ./data.

Please also download the pretrained model from here (BaiduNetDisk, OneDrive, password: gk8y).

Evaluation on YCB-V

python test.py --use_gpu --filelist=FILELIST --out_dir=./Result --test_mode=YCB-Video --model_path=MODEL_PATH

Example:

python test.py --use_gpu --filelist=ycb-video-testlist.txt --out_dir=./Result --test_mode=YCB-Video --model_path=dgecn.pth

Visualize depth predictions

python depth_vis.py --image_path=IMAGE_PATH --model_path=MODEL_PATH --ext==FILE_EXT

Example:

python depth_vis.py --image_path=./assert/test_0.png --model_path=dgecn.pth --ext=png

This pretrained model is trained on video 0000 ~ 0010， the predictions on other videos may not good. We will release our pretrained model on all videos in the full code.

Citation

If you find this useful in your research, please consider citing our paper.

@InProceedings{Cao_2022_CVPR,
    author    = {Cao, Tuo and Luo, Fei and Fu, Yanping and Zhang, Wenxiao and Zheng, Shengjie and Xiao, Chunxia},
    title     = {DGECN: A Depth-Guided Edge Convolutional Network for End-to-End 6D Pose Estimation},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month     = {June},
    year      = {2022},
    pages     = {3783-3792}
}

References

• [1] Sida Peng, Yuan Liu, Qixing Huang, Hujun Bao, and Xiaowei Zhou. PVNET: Pixel-wise voting network for 6dof pose estimation.
• [2] Yinlin Hu, Joachim Hugonot, Pascal Fua, and Mathieu Salz- mann. Segmentation-Driven 6D Object Pose Estimation..
• [3] Clément Godard, Oisin Mac Aodha, Michael Firman and Gabriel J. Brostow. Digging into Self-Supervised Monocular Depth Prediction.
• [4] Wang, Yue; Sun, Yongbin; Liu, Ziwei; Sarma, Sanjay E.; Bronstein, Michael M.; Solomon, Justin M. Dynamic Graph CNN for Learning on Point Clouds.
• [4] Gu Wang, Fabian Manhardt, Federico Tombari, and Xi- angyang Ji. GDR-net: Geometry-guided direct regression network for monocular 6d object pose estimation.