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

2025-02-04

AI Summary

To address the heavy reliance on large-scale real-world 3D annotations for category-level 9-DoF (6D pose + 3D size) object estimation in AR and robotic manipulation, this paper proposes the first generative approach based on Denoising Diffusion Implicit Models (DDIM), requiring neither 3D shape priors nor real training data. Our method takes only synthetic rendered images as input and jointly decodes pose and size via end-to-end differentiable rendering to enforce observation consistency—fully prior-free. We introduce a three-step fast sampling strategy that balances accuracy and efficiency. Evaluated on two benchmark datasets and a real robotic arm platform, our method achieves state-of-the-art domain generalization performance. With only three diffusion iterations at inference time, it enables near-real-time operation while maintaining high fidelity.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Intelligent Robots: State Estimation
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Security and Privacy: Large-scale security measurements

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.

Problem

• Addressing domain generalization in 9-DoF pose estimation.
• Training model on synthetic data for real-world application.
• Achieving real-time performance with diffusion models.

Innovation

• Diffusion-based domain-generalized object pose estimation
• Training exclusively on synthetic rendered data
• Denoising Diffusion Implicit Model for real-time performance