@article{
 title = {GDR-Net: Geometry-Guided Direct Regression Network for Monocular 6D Object Pose Estimation},
 type = {article},
 year = {2021},
 keywords = {Direct regression,Object pose estimation,Prior geometry},
 pages = {6241-6246},
 volume = {2022-July},
 websites = {https://arxiv.org/abs/2102.12145v3},
 month = {2},
 publisher = {IEEE Computer Society},
 day = {24},
 id = {61717ad9-a583-3bd1-ba3f-b18bdcbe8f62},
 created = {2024-08-20T07:23:52.193Z},
 accessed = {2024-08-20},
 file_attached = {true},
 profile_id = {f1f70cad-e32d-3de2-a3c0-be1736cb88be},
 group_id = {5ec9cc91-a5d6-3de5-82f3-3ef3d98a89c1},
 last_modified = {2024-08-20T07:23:57.137Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {false},
 hidden = {false},
 folder_uuids = {df28411a-ed7f-4991-8358-d39685eb4bf0},
 private_publication = {false},
 abstract = {6D pose estimation from a single RGB image is a fundamental task in computer vision. The current top-performing deep learning-based methods rely on an indirect strategy, i.e., first establishing 2D-3D correspondences between the coordinates in the image plane and object coordinate system, and then applying a variant of the P$n$P/RANSAC algorithm. However, this two-stage pipeline is not end-to-end trainable, thus is hard to be employed for many tasks requiring differentiable poses. On the other hand, methods based on direct regression are currently inferior to geometry-based methods. In this work, we perform an in-depth investigation on both direct and indirect methods, and propose a simple yet effective Geometry-guided Direct Regression Network (GDR-Net) to learn the 6D pose in an end-to-end manner from dense correspondence-based intermediate geometric representations. Extensive experiments show that our approach remarkably outperforms state-of-the-art methods on LM, LM-O and YCB-V datasets. Code is available at https://git.io/GDR-Net.},
 bibtype = {article},
 author = {Liu, Chongpei and Sun, Wei and Zhang, Keyi and Liu, Jian and Zhang, Xing and Fan, Shimeng},
 doi = {10.23919/CCC55666.2022.9901912},
 journal = {Chinese Control Conference, CCC}
}

{"_id":"kYKTTQNKdMeirW8Zx","bibbaseid":"liu-sun-zhang-liu-zhang-fan-gdrnetgeometryguideddirectregressionnetworkformonocular6dobjectposeestimation-2021","author_short":["Liu, C.","Sun, W.","Zhang, K.","Liu, J.","Zhang, X.","Fan, S."],"bibdata":{"title":"GDR-Net: Geometry-Guided Direct Regression Network for Monocular 6D Object Pose Estimation","type":"article","year":"2021","keywords":"Direct regression,Object pose estimation,Prior geometry","pages":"6241-6246","volume":"2022-July","websites":"https://arxiv.org/abs/2102.12145v3","month":"2","publisher":"IEEE Computer Society","day":"24","id":"61717ad9-a583-3bd1-ba3f-b18bdcbe8f62","created":"2024-08-20T07:23:52.193Z","accessed":"2024-08-20","file_attached":"true","profile_id":"f1f70cad-e32d-3de2-a3c0-be1736cb88be","group_id":"5ec9cc91-a5d6-3de5-82f3-3ef3d98a89c1","last_modified":"2024-08-20T07:23:57.137Z","read":false,"starred":false,"authored":false,"confirmed":false,"hidden":false,"folder_uuids":"df28411a-ed7f-4991-8358-d39685eb4bf0","private_publication":false,"abstract":"6D pose estimation from a single RGB image is a fundamental task in computer vision. The current top-performing deep learning-based methods rely on an indirect strategy, i.e., first establishing 2D-3D correspondences between the coordinates in the image plane and object coordinate system, and then applying a variant of the P$n$P/RANSAC algorithm. However, this two-stage pipeline is not end-to-end trainable, thus is hard to be employed for many tasks requiring differentiable poses. On the other hand, methods based on direct regression are currently inferior to geometry-based methods. In this work, we perform an in-depth investigation on both direct and indirect methods, and propose a simple yet effective Geometry-guided Direct Regression Network (GDR-Net) to learn the 6D pose in an end-to-end manner from dense correspondence-based intermediate geometric representations. Extensive experiments show that our approach remarkably outperforms state-of-the-art methods on LM, LM-O and YCB-V datasets. Code is available at https://git.io/GDR-Net.","bibtype":"article","author":"Liu, Chongpei and Sun, Wei and Zhang, Keyi and Liu, Jian and Zhang, Xing and Fan, Shimeng","doi":"10.23919/CCC55666.2022.9901912","journal":"Chinese Control Conference, CCC","bibtex":"@article{\n title = {GDR-Net: Geometry-Guided Direct Regression Network for Monocular 6D Object Pose Estimation},\n type = {article},\n year = {2021},\n keywords = {Direct regression,Object pose estimation,Prior geometry},\n pages = {6241-6246},\n volume = {2022-July},\n websites = {https://arxiv.org/abs/2102.12145v3},\n month = {2},\n publisher = {IEEE Computer Society},\n day = {24},\n id = {61717ad9-a583-3bd1-ba3f-b18bdcbe8f62},\n created = {2024-08-20T07:23:52.193Z},\n accessed = {2024-08-20},\n file_attached = {true},\n profile_id = {f1f70cad-e32d-3de2-a3c0-be1736cb88be},\n group_id = {5ec9cc91-a5d6-3de5-82f3-3ef3d98a89c1},\n last_modified = {2024-08-20T07:23:57.137Z},\n read = {false},\n starred = {false},\n authored = {false},\n confirmed = {false},\n hidden = {false},\n folder_uuids = {df28411a-ed7f-4991-8358-d39685eb4bf0},\n private_publication = {false},\n abstract = {6D pose estimation from a single RGB image is a fundamental task in computer vision. The current top-performing deep learning-based methods rely on an indirect strategy, i.e., first establishing 2D-3D correspondences between the coordinates in the image plane and object coordinate system, and then applying a variant of the P$n$P/RANSAC algorithm. However, this two-stage pipeline is not end-to-end trainable, thus is hard to be employed for many tasks requiring differentiable poses. On the other hand, methods based on direct regression are currently inferior to geometry-based methods. In this work, we perform an in-depth investigation on both direct and indirect methods, and propose a simple yet effective Geometry-guided Direct Regression Network (GDR-Net) to learn the 6D pose in an end-to-end manner from dense correspondence-based intermediate geometric representations. Extensive experiments show that our approach remarkably outperforms state-of-the-art methods on LM, LM-O and YCB-V datasets. Code is available at https://git.io/GDR-Net.},\n bibtype = {article},\n author = {Liu, Chongpei and Sun, Wei and Zhang, Keyi and Liu, Jian and Zhang, Xing and Fan, Shimeng},\n doi = {10.23919/CCC55666.2022.9901912},\n journal = {Chinese Control Conference, CCC}\n}","author_short":["Liu, C.","Sun, W.","Zhang, K.","Liu, J.","Zhang, X.","Fan, S."],"urls":{"Paper":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c/file/459b058f-1236-047f-2a7a-947c67665e7a/full_text.pdf.pdf","Website":"https://arxiv.org/abs/2102.12145v3"},"biburl":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c","bibbaseid":"liu-sun-zhang-liu-zhang-fan-gdrnetgeometryguideddirectregressionnetworkformonocular6dobjectposeestimation-2021","role":"author","keyword":["Direct regression","Object pose estimation","Prior geometry"],"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"article","biburl":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c","dataSources":["2252seNhipfTmjEBQ"],"keywords":["direct regression","object pose estimation","prior geometry"],"search_terms":["gdr","net","geometry","guided","direct","regression","network","monocular","object","pose","estimation","liu","sun","zhang","liu","zhang","fan"],"title":"GDR-Net: Geometry-Guided Direct Regression Network for Monocular 6D Object Pose Estimation","year":2021}