Home > Research > Publications & Outputs > Diff9D

Associated organisational unit

Electronic data

  • Diff9D

    Accepted author manuscript, 8.44 MB, PDF document

    Available under license: CC BY: Creative Commons Attribution 4.0 International License

Links

Text available via DOI:

View graph of relations

Diff9D: Diffusion-Based Domain-Generalized Category-Level 9-DoF Object Pose Estimation

Research output: Contribution to Journal/MagazineJournal articlepeer-review

E-pub ahead of print

Standard

Diff9D: Diffusion-Based Domain-Generalized Category-Level 9-DoF Object Pose Estimation. / Liu, Jian; Sun, Wei; Yang, Hui et al.
In: IEEE Transactions on Pattern Analysis and Machine Intelligence, 18.03.2025, p. 1-17.

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Harvard

Liu, J, Sun, W, Yang, H, Deng, P, Liu, C, Sebe, N, Rahmani, H & Mian, A 2025, 'Diff9D: Diffusion-Based Domain-Generalized Category-Level 9-DoF Object Pose Estimation', IEEE Transactions on Pattern Analysis and Machine Intelligence, pp. 1-17. https://doi.org/10.1109/tpami.2025.3552132

APA

Liu, J., Sun, W., Yang, H., Deng, P., Liu, C., Sebe, N., Rahmani, H., & Mian, A. (2025). Diff9D: Diffusion-Based Domain-Generalized Category-Level 9-DoF Object Pose Estimation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1-17. Advance online publication. https://doi.org/10.1109/tpami.2025.3552132

Vancouver

Liu J, Sun W, Yang H, Deng P, Liu C, Sebe N et al. Diff9D: Diffusion-Based Domain-Generalized Category-Level 9-DoF Object Pose Estimation. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2025 Mar 18;1-17. Epub 2025 Mar 18. doi: 10.1109/tpami.2025.3552132

Author

Liu, Jian ; Sun, Wei ; Yang, Hui et al. / Diff9D : Diffusion-Based Domain-Generalized Category-Level 9-DoF Object Pose Estimation. In: IEEE Transactions on Pattern Analysis and Machine Intelligence. 2025 ; pp. 1-17.

Bibtex

@article{0e75710a80b3403884c8db394fda4fd6,
title = "Diff9D: Diffusion-Based Domain-Generalized Category-Level 9-DoF Object Pose Estimation",
abstract = "Nine-degrees-of-freedom (9-DoF) object pose and size estimation is crucial for enabling augmented reality and robotic manipulation. Category-level methods have received extensive research attention due to their potential for generalization to intra-class unknown objects. However, these methods require manual collection and labeling of large-scale real-world training data. To address this problem, we introduce a diffusion-based paradigm for domain-generalized category-level 9-DoF object pose estimation. Our motivation is to leverage the latent generalization ability of the diffusion model to address the domain generalization challenge in object pose estimation. This entails training the model exclusively on rendered synthetic data to achieve generalization to real-world scenes. We propose an effective diffusion model to redefine 9-DoF object pose estimation from a generative perspective. Our model does not require any 3D shape priors during training or inference. By employing the Denoising Diffusion Implicit Model, we demonstrate that the reverse diffusion process can be executed in as few as 3 steps, achieving near real-time performance. Finally, we design a robotic grasping system comprising both hardware and software components. Through comprehensive experiments on two benchmark datasets and the real-world robotic system, we show that our method achieves state-of-the-art domain generalization performance. Our code will be made public at https://github.com/CNJianLiu/Diff9D.",
author = "Jian Liu and Wei Sun and Hui Yang and Pengchao Deng and Chongpei Liu and Nicu Sebe and Hossein Rahmani and Ajmal Mian",
year = "2025",
month = mar,
day = "18",
doi = "10.1109/tpami.2025.3552132",
language = "English",
pages = "1--17",
journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
issn = "0162-8828",
publisher = "IEEE Computer Society",

}

RIS

TY - JOUR

T1 - Diff9D

T2 - Diffusion-Based Domain-Generalized Category-Level 9-DoF Object Pose Estimation

AU - Liu, Jian

AU - Sun, Wei

AU - Yang, Hui

AU - Deng, Pengchao

AU - Liu, Chongpei

AU - Sebe, Nicu

AU - Rahmani, Hossein

AU - Mian, Ajmal

PY - 2025/3/18

Y1 - 2025/3/18

N2 - Nine-degrees-of-freedom (9-DoF) object pose and size estimation is crucial for enabling augmented reality and robotic manipulation. Category-level methods have received extensive research attention due to their potential for generalization to intra-class unknown objects. However, these methods require manual collection and labeling of large-scale real-world training data. To address this problem, we introduce a diffusion-based paradigm for domain-generalized category-level 9-DoF object pose estimation. Our motivation is to leverage the latent generalization ability of the diffusion model to address the domain generalization challenge in object pose estimation. This entails training the model exclusively on rendered synthetic data to achieve generalization to real-world scenes. We propose an effective diffusion model to redefine 9-DoF object pose estimation from a generative perspective. Our model does not require any 3D shape priors during training or inference. By employing the Denoising Diffusion Implicit Model, we demonstrate that the reverse diffusion process can be executed in as few as 3 steps, achieving near real-time performance. Finally, we design a robotic grasping system comprising both hardware and software components. Through comprehensive experiments on two benchmark datasets and the real-world robotic system, we show that our method achieves state-of-the-art domain generalization performance. Our code will be made public at https://github.com/CNJianLiu/Diff9D.

AB - Nine-degrees-of-freedom (9-DoF) object pose and size estimation is crucial for enabling augmented reality and robotic manipulation. Category-level methods have received extensive research attention due to their potential for generalization to intra-class unknown objects. However, these methods require manual collection and labeling of large-scale real-world training data. To address this problem, we introduce a diffusion-based paradigm for domain-generalized category-level 9-DoF object pose estimation. Our motivation is to leverage the latent generalization ability of the diffusion model to address the domain generalization challenge in object pose estimation. This entails training the model exclusively on rendered synthetic data to achieve generalization to real-world scenes. We propose an effective diffusion model to redefine 9-DoF object pose estimation from a generative perspective. Our model does not require any 3D shape priors during training or inference. By employing the Denoising Diffusion Implicit Model, we demonstrate that the reverse diffusion process can be executed in as few as 3 steps, achieving near real-time performance. Finally, we design a robotic grasping system comprising both hardware and software components. Through comprehensive experiments on two benchmark datasets and the real-world robotic system, we show that our method achieves state-of-the-art domain generalization performance. Our code will be made public at https://github.com/CNJianLiu/Diff9D.

U2 - 10.1109/tpami.2025.3552132

DO - 10.1109/tpami.2025.3552132

M3 - Journal article

SP - 1

EP - 17

JO - IEEE Transactions on Pattern Analysis and Machine Intelligence

JF - IEEE Transactions on Pattern Analysis and Machine Intelligence

SN - 0162-8828

ER -