Stable-Sim2Real: Exploring Simulation of Real-Captured 3D Data with Two-Stage Depth Diffusion

• URL: http://arxiv.org/abs/2507.23483v1
• Date: Thu, 31 Jul 2025 12:08:16 GMT
• Title: Stable-Sim2Real: Exploring Simulation of Real-Captured 3D Data with Two-Stage Depth Diffusion
• Authors: Mutian Xu, Chongjie Ye, Haolin Liu, Yushuang Wu, Jiahao Chang, Xiaoguang Han,
• Abstract summary: 3D data simulation aims to bridge the gap between simulated and real-captured 3D data.<n>Most 3D data simulation methods inject predefined physical priors but struggle to capture the full complexity of real data.<n>This work explores a new solution path, called Stable-Sim2Real, based on a novel two-stage depth diffusion model.
• Score: 16.720863475636328
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: 3D data simulation aims to bridge the gap between simulated and real-captured 3D data, which is a fundamental problem for real-world 3D visual tasks. Most 3D data simulation methods inject predefined physical priors but struggle to capture the full complexity of real data. An optimal approach involves learning an implicit mapping from synthetic to realistic data in a data-driven manner, but progress in this solution has met stagnation in recent studies. This work explores a new solution path of data-driven 3D simulation, called Stable-Sim2Real, based on a novel two-stage depth diffusion model. The initial stage finetunes Stable-Diffusion to generate the residual between the real and synthetic paired depth, producing a stable but coarse depth, where some local regions may deviate from realistic patterns. To enhance this, both the synthetic and initial output depth are fed into a second-stage diffusion, where diffusion loss is adjusted to prioritize these distinct areas identified by a 3D discriminator. We provide a new benchmark scheme to evaluate 3D data simulation methods. Extensive experiments show that training the network with the 3D simulated data derived from our method significantly enhances performance in real-world 3D visual tasks. Moreover, the evaluation demonstrates the high similarity between our 3D simulated data and real-captured patterns. Project page: https://mutianxu.github.io/stable-sim2real/.

Related papers

• R3D2: Realistic 3D Asset Insertion via Diffusion for Autonomous Driving Simulation [78.26308457952636]
  This paper introduces R3D2, a lightweight, one-step diffusion model designed to overcome limitations in autonomous driving simulation.<n>It enables realistic insertion of complete 3D assets into existing scenes by generating plausible rendering effects-such as shadows and consistent lighting-in real time.<n>We show that R3D2 significantly enhances the realism of inserted assets, enabling use-cases like text-to-3D asset insertion and cross-scene/dataset object transfer.
  arXiv  Detail & Related papers  (2025-06-09T14:50:19Z)
• Enhancing Generalizability of Representation Learning for Data-Efficient 3D Scene Understanding [50.448520056844885]
  We propose a generative Bayesian network to produce diverse synthetic scenes with real-world patterns. A series of experiments robustly display our method's consistent superiority over existing state-of-the-art pre-training approaches.
  arXiv  Detail & Related papers  (2024-06-17T07:43:53Z)
• DIRECT-3D: Learning Direct Text-to-3D Generation on Massive Noisy 3D Data [50.164670363633704]
  We present DIRECT-3D, a diffusion-based 3D generative model for creating high-quality 3D assets from text prompts. Our model is directly trained on extensive noisy and unaligned in-the-wild' 3D assets. We achieve state-of-the-art performance in both single-class generation and text-to-3D generation.
  arXiv  Detail & Related papers  (2024-06-06T17:58:15Z)
• A New Benchmark: On the Utility of Synthetic Data with Blender for Bare Supervised Learning and Downstream Domain Adaptation [42.2398858786125]
  Deep learning in computer vision has achieved great success with the price of large-scale labeled training data. The uncontrollable data collection process produces non-IID training and test data, where undesired duplication may exist. To circumvent them, an alternative is to generate synthetic data via 3D rendering with domain randomization.
  arXiv  Detail & Related papers  (2023-03-16T09:03:52Z)
• DreamFusion: Text-to-3D using 2D Diffusion [52.52529213936283]
  Recent breakthroughs in text-to-image synthesis have been driven by diffusion models trained on billions of image-text pairs. In this work, we circumvent these limitations by using a pretrained 2D text-to-image diffusion model to perform text-to-3D synthesis. Our approach requires no 3D training data and no modifications to the image diffusion model, demonstrating the effectiveness of pretrained image diffusion models as priors.
  arXiv  Detail & Related papers  (2022-09-29T17:50:40Z)
• Self-supervised Human Mesh Recovery with Cross-Representation Alignment [20.69546341109787]
  Self-supervised human mesh recovery methods have poor generalizability due to limited availability and diversity of 3D-annotated benchmark datasets. We propose cross-representation alignment utilizing the complementary information from the robust but sparse representation (2D keypoints) This adaptive cross-representation alignment explicitly learns from the deviations and captures complementary information: richness from sparse representation and robustness from dense representation.
  arXiv  Detail & Related papers  (2022-09-10T04:47:20Z)
• Learning Dense Correspondence from Synthetic Environments [27.841736037738286]
  Existing methods map manually labelled human pixels in real 2D images onto the 3D surface, which is prone to human error. We propose to solve the problem of data scarcity by training 2D-3D human mapping algorithms using automatically generated synthetic data.
  arXiv  Detail & Related papers  (2022-03-24T08:13:26Z)
• RandomRooms: Unsupervised Pre-training from Synthetic Shapes and Randomized Layouts for 3D Object Detection [138.2892824662943]
  A promising solution is to make better use of the synthetic dataset, which consists of CAD object models, to boost the learning on real datasets. Recent work on 3D pre-training exhibits failure when transfer features learned on synthetic objects to other real-world applications. In this work, we put forward a new method called RandomRooms to accomplish this objective.
  arXiv  Detail & Related papers  (2021-08-17T17:56:12Z)
• Secrets of 3D Implicit Object Shape Reconstruction in the Wild [92.5554695397653]
  Reconstructing high-fidelity 3D objects from sparse, partial observation is crucial for various applications in computer vision, robotics, and graphics. Recent neural implicit modeling methods show promising results on synthetic or dense datasets. But, they perform poorly on real-world data that is sparse and noisy. This paper analyzes the root cause of such deficient performance of a popular neural implicit model.
  arXiv  Detail & Related papers  (2021-01-18T03:24:48Z)
• Exploring the Capabilities and Limits of 3D Monocular Object Detection -- A Study on Simulation and Real World Data [0.0]
  3D object detection based on monocular camera data is key enabler for autonomous driving. Recent deep learning methods show promising results to recover depth information from single images. In this paper, we evaluate the performance of a 3D object detection pipeline which is parameterizable with different depth estimation configurations.
  arXiv  Detail & Related papers  (2020-05-15T09:05:17Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.