Decompositional Neural Scene Reconstruction with Generative Diffusion Prior
- URL: http://arxiv.org/abs/2503.14830v1
- Date: Wed, 19 Mar 2025 02:11:31 GMT
- Title: Decompositional Neural Scene Reconstruction with Generative Diffusion Prior
- Authors: Junfeng Ni, Yu Liu, Ruijie Lu, Zirui Zhou, Song-Chun Zhu, Yixin Chen, Siyuan Huang,
- Abstract summary: Decompositional reconstruction of 3D scenes, with complete shapes and detailed texture, is intriguing for downstream applications.<n>Recent approaches incorporate semantic or geometric regularization to address this issue, but they suffer significant degradation in underconstrained areas.<n>We propose DP-Recon, which employs diffusion priors in the form of Score Distillation Sampling (SDS) to optimize the neural representation of each individual object under novel views.
- Score: 64.71091831762214
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: Decompositional reconstruction of 3D scenes, with complete shapes and detailed texture of all objects within, is intriguing for downstream applications but remains challenging, particularly with sparse views as input. Recent approaches incorporate semantic or geometric regularization to address this issue, but they suffer significant degradation in underconstrained areas and fail to recover occluded regions. We argue that the key to solving this problem lies in supplementing missing information for these areas. To this end, we propose DP-Recon, which employs diffusion priors in the form of Score Distillation Sampling (SDS) to optimize the neural representation of each individual object under novel views. This provides additional information for the underconstrained areas, but directly incorporating diffusion prior raises potential conflicts between the reconstruction and generative guidance. Therefore, we further introduce a visibility-guided approach to dynamically adjust the per-pixel SDS loss weights. Together these components enhance both geometry and appearance recovery while remaining faithful to input images. Extensive experiments across Replica and ScanNet++ demonstrate that our method significantly outperforms SOTA methods. Notably, it achieves better object reconstruction under 10 views than the baselines under 100 views. Our method enables seamless text-based editing for geometry and appearance through SDS optimization and produces decomposed object meshes with detailed UV maps that support photorealistic Visual effects (VFX) editing. The project page is available at https://dp-recon.github.io/.
Related papers
- How to Use Diffusion Priors under Sparse Views? [29.738350228085928]
Inline Prior Guided Score Matching is proposed to provide visual supervision over sparse views in 3D reconstruction.<n>We show that our method achieves state-of-the-art reconstruction quality.
arXiv Detail & Related papers (2024-12-03T07:31:54Z) - LM-Gaussian: Boost Sparse-view 3D Gaussian Splatting with Large Model Priors [34.91966359570867]
sparse-view reconstruction is inherently ill-posed and under-constrained.
We introduce LM-Gaussian, a method capable of generating high-quality reconstructions from a limited number of images.
Our approach significantly reduces the data acquisition requirements compared to previous 3DGS methods.
arXiv Detail & Related papers (2024-09-05T12:09:02Z) - Towards Cross-View-Consistent Self-Supervised Surround Depth Estimation [9.569646683579899]
Self-Supervised Surround Depth Estimation from consecutive images offers an economical alternative.
Previous SSSDE methods have proposed different mechanisms to fuse information across images, but few of them explicitly consider the cross-view constraints.
This paper proposes an efficient and consistent pose estimation design and two loss functions to enhance cross-view consistency for SSSDE.
arXiv Detail & Related papers (2024-07-04T16:29:05Z) - SMORE: Simulataneous Map and Object REconstruction [66.66729715211642]
We present a method for dynamic surface reconstruction of large-scale urban scenes from LiDAR.<n>We take a holistic perspective and optimize a compositional model of a dynamic scene that decomposes the world into rigidly-moving objects and the background.
arXiv Detail & Related papers (2024-06-19T23:53:31Z) - Total-Decom: Decomposed 3D Scene Reconstruction with Minimal Interaction [51.3632308129838]
We present Total-Decom, a novel method for decomposed 3D reconstruction with minimal human interaction.
Our approach seamlessly integrates the Segment Anything Model (SAM) with hybrid implicit-explicit neural surface representations and a mesh-based region-growing technique for accurate 3D object decomposition.
We extensively evaluate our method on benchmark datasets and demonstrate its potential for downstream applications, such as animation and scene editing.
arXiv Detail & Related papers (2024-03-28T11:12:33Z) - AugUndo: Scaling Up Augmentations for Monocular Depth Completion and Estimation [51.143540967290114]
We propose a method that unlocks a wide range of previously-infeasible geometric augmentations for unsupervised depth computation and estimation.
This is achieved by reversing, or undo''-ing, geometric transformations to the coordinates of the output depth, warping the depth map back to the original reference frame.
arXiv Detail & Related papers (2023-10-15T05:15:45Z) - FrozenRecon: Pose-free 3D Scene Reconstruction with Frozen Depth Models [67.96827539201071]
We propose a novel test-time optimization approach for 3D scene reconstruction.
Our method achieves state-of-the-art cross-dataset reconstruction on five zero-shot testing datasets.
arXiv Detail & Related papers (2023-08-10T17:55:02Z) - ERF: Explicit Radiance Field Reconstruction From Scratch [12.254150867994163]
We propose a novel explicit dense 3D reconstruction approach that processes a set of images of a scene with sensor poses and calibrations and estimates a photo-real digital model.
One of the key innovations is that the underlying volumetric representation is completely explicit.
We show that our method is general and practical. It does not require a highly controlled lab setup for capturing, but allows for reconstructing scenes with a vast variety of objects.
arXiv Detail & Related papers (2022-02-28T19:37:12Z) - 3D Dense Geometry-Guided Facial Expression Synthesis by Adversarial
Learning [54.24887282693925]
We propose a novel framework to exploit 3D dense (depth and surface normals) information for expression manipulation.
We use an off-the-shelf state-of-the-art 3D reconstruction model to estimate the depth and create a large-scale RGB-Depth dataset.
Our experiments demonstrate that the proposed method outperforms the competitive baseline and existing arts by a large margin.
arXiv Detail & Related papers (2020-09-30T17:12:35Z)
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.