2024-04-03 更新

Multi-Level Neural Scene Graphs for Dynamic Urban Environments

Authors:Tobias Fischer, Lorenzo Porzi, Samuel Rota Bulò, Marc Pollefeys, Peter Kontschieder

We estimate the radiance field of large-scale dynamic areas from multiple vehicle captures under varying environmental conditions. Previous works in this domain are either restricted to static environments, do not scale to more than a single short video, or struggle to separately represent dynamic object instances. To this end, we present a novel, decomposable radiance field approach for dynamic urban environments. We propose a multi-level neural scene graph representation that scales to thousands of images from dozens of sequences with hundreds of fast-moving objects. To enable efficient training and rendering of our representation, we develop a fast composite ray sampling and rendering scheme. To test our approach in urban driving scenarios, we introduce a new, novel view synthesis benchmark. We show that our approach outperforms prior art by a significant margin on both established and our proposed benchmark while being faster in training and rendering.
PDF CVPR 2024. Project page is available at


3DGSR: Implicit Surface Reconstruction with 3D Gaussian Splatting

Authors:Xiaoyang Lyu, Yang-Tian Sun, Yi-Hua Huang, Xiuzhe Wu, Ziyi Yang, Yilun Chen, Jiangmiao Pang, Xiaojuan Qi

In this paper, we present an implicit surface reconstruction method with 3D Gaussian Splatting (3DGS), namely 3DGSR, that allows for accurate 3D reconstruction with intricate details while inheriting the high efficiency and rendering quality of 3DGS. The key insight is incorporating an implicit signed distance field (SDF) within 3D Gaussians to enable them to be aligned and jointly optimized. First, we introduce a differentiable SDF-to-opacity transformation function that converts SDF values into corresponding Gaussians’ opacities. This function connects the SDF and 3D Gaussians, allowing for unified optimization and enforcing surface constraints on the 3D Gaussians. During learning, optimizing the 3D Gaussians provides supervisory signals for SDF learning, enabling the reconstruction of intricate details. However, this only provides sparse supervisory signals to the SDF at locations occupied by Gaussians, which is insufficient for learning a continuous SDF. Then, to address this limitation, we incorporate volumetric rendering and align the rendered geometric attributes (depth, normal) with those derived from 3D Gaussians. This consistency regularization introduces supervisory signals to locations not covered by discrete 3D Gaussians, effectively eliminating redundant surfaces outside the Gaussian sampling range. Our extensive experimental results demonstrate that our 3DGSR method enables high-quality 3D surface reconstruction while preserving the efficiency and rendering quality of 3DGS. Besides, our method competes favorably with leading surface reconstruction techniques while offering a more efficient learning process and much better rendering qualities. The code will be available at


Knowledge NeRF: Few-shot Novel View Synthesis for Dynamic Articulated Objects

Authors:Wenxiao Cai, Xinyue Leiınst, Xinyu He, Junming Leo Chen, Yangang Wang

We present Knowledge NeRF to synthesize novel views for dynamic scenes.Reconstructing dynamic 3D scenes from few sparse views and rendering them from arbitrary perspectives is a challenging problem with applications in various domains. Previous dynamic NeRF methods learn the deformation of articulated objects from monocular videos. However, qualities of their reconstructed scenes are limited.To clearly reconstruct dynamic scenes, we propose a new framework by considering two frames at a time.We pretrain a NeRF model for an articulated object.When articulated objects moves, Knowledge NeRF learns to generate novel views at the new state by incorporating past knowledge in the pretrained NeRF model with minimal observations in the present state. We propose a projection module to adapt NeRF for dynamic scenes, learning the correspondence between pretrained knowledge base and current states. Experimental results demonstrate the effectiveness of our method in reconstructing dynamic 3D scenes with 5 input images in one state. Knowledge NeRF is a new pipeline and promising solution for novel view synthesis in dynamic articulated objects. The data and implementation are publicly available at


OmniLocalRF: Omnidirectional Local Radiance Fields from Dynamic Videos

Authors:Dongyoung Choi, Hyeonjoong Jang, Min H. Kim

Omnidirectional cameras are extensively used in various applications to provide a wide field of vision. However, they face a challenge in synthesizing novel views due to the inevitable presence of dynamic objects, including the photographer, in their wide field of view. In this paper, we introduce a new approach called Omnidirectional Local Radiance Fields (OmniLocalRF) that can render static-only scene views, removing and inpainting dynamic objects simultaneously. Our approach combines the principles of local radiance fields with the bidirectional optimization of omnidirectional rays. Our input is an omnidirectional video, and we evaluate the mutual observations of the entire angle between the previous and current frames. To reduce ghosting artifacts of dynamic objects and inpaint occlusions, we devise a multi-resolution motion mask prediction module. Unlike existing methods that primarily separate dynamic components through the temporal domain, our method uses multi-resolution neural feature planes for precise segmentation, which is more suitable for long 360-degree videos. Our experiments validate that OmniLocalRF outperforms existing methods in both qualitative and quantitative metrics, especially in scenarios with complex real-world scenes. In particular, our approach eliminates the need for manual interaction, such as drawing motion masks by hand and additional pose estimation, making it a highly effective and efficient solution.


DPA-Net: Structured 3D Abstraction from Sparse Views via Differentiable Primitive Assembly

Authors:Fenggen Yu, Yiming Qian, Xu Zhang, Francisca Gil-Ureta, Brian Jackson, Eric Bennett, Hao Zhang

We present a differentiable rendering framework to learn structured 3D abstractions in the form of primitive assemblies from sparse RGB images capturing a 3D object. By leveraging differentiable volume rendering, our method does not require 3D supervision. Architecturally, our network follows the general pipeline of an image-conditioned neural radiance field (NeRF) exemplified by pixelNeRF for color prediction. As our core contribution, we introduce differential primitive assembly (DPA) into NeRF to output a 3D occupancy field in place of density prediction, where the predicted occupancies serve as opacity values for volume rendering. Our network, coined DPA-Net, produces a union of convexes, each as an intersection of convex quadric primitives, to approximate the target 3D object, subject to an abstraction loss and a masking loss, both defined in the image space upon volume rendering. With test-time adaptation and additional sampling and loss designs aimed at improving the accuracy and compactness of the obtained assemblies, our method demonstrates superior performance over state-of-the-art alternatives for 3D primitive abstraction from sparse views.
PDF 14 pages


Marrying NeRF with Feature Matching for One-step Pose Estimation

Authors:Ronghan Chen, Yang Cong, Yu Ren

Given the image collection of an object, we aim at building a real-time image-based pose estimation method, which requires neither its CAD model nor hours of object-specific training. Recent NeRF-based methods provide a promising solution by directly optimizing the pose from pixel loss between rendered and target images. However, during inference, they require long converging time, and suffer from local minima, making them impractical for real-time robot applications. We aim at solving this problem by marrying image matching with NeRF. With 2D matches and depth rendered by NeRF, we directly solve the pose in one step by building 2D-3D correspondences between target and initial view, thus allowing for real-time prediction. Moreover, to improve the accuracy of 2D-3D correspondences, we propose a 3D consistent point mining strategy, which effectively discards unfaithful points reconstruted by NeRF. Moreover, current NeRF-based methods naively optimizing pixel loss fail at occluded images. Thus, we further propose a 2D matches based sampling strategy to preclude the occluded area. Experimental results on representative datasets prove that our method outperforms state-of-the-art methods, and improves inference efficiency by 90x, achieving real-time prediction at 6 FPS.
PDF ICRA, 2024. Video


MM3DGS SLAM: Multi-modal 3D Gaussian Splatting for SLAM Using Vision, Depth, and Inertial Measurements

Authors:Lisong C. Sun, Neel P. Bhatt, Jonathan C. Liu, Zhiwen Fan, Zhangyang Wang, Todd E. Humphreys, Ufuk Topcu

Simultaneous localization and mapping is essential for position tracking and scene understanding. 3D Gaussian-based map representations enable photorealistic reconstruction and real-time rendering of scenes using multiple posed cameras. We show for the first time that using 3D Gaussians for map representation with unposed camera images and inertial measurements can enable accurate SLAM. Our method, MM3DGS, addresses the limitations of prior neural radiance field-based representations by enabling faster rendering, scale awareness, and improved trajectory tracking. Our framework enables keyframe-based mapping and tracking utilizing loss functions that incorporate relative pose transformations from pre-integrated inertial measurements, depth estimates, and measures of photometric rendering quality. We also release a multi-modal dataset, UT-MM, collected from a mobile robot equipped with a camera and an inertial measurement unit. Experimental evaluation on several scenes from the dataset shows that MM3DGS achieves 3x improvement in tracking and 5% improvement in photometric rendering quality compared to the current 3DGS SLAM state-of-the-art, while allowing real-time rendering of a high-resolution dense 3D map. Project Webpage:
PDF Project Webpage:


CityGaussian: Real-time High-quality Large-Scale Scene Rendering with Gaussians

Authors:Yang Liu, He Guan, Chuanchen Luo, Lue Fan, Junran Peng, Zhaoxiang Zhang

The advancement of real-time 3D scene reconstruction and novel view synthesis has been significantly propelled by 3D Gaussian Splatting (3DGS). However, effectively training large-scale 3DGS and rendering it in real-time across various scales remains challenging. This paper introduces CityGaussian (CityGS), which employs a novel divide-and-conquer training approach and Level-of-Detail (LoD) strategy for efficient large-scale 3DGS training and rendering. Specifically, the global scene prior and adaptive training data selection enables efficient training and seamless fusion. Based on fused Gaussian primitives, we generate different detail levels through compression, and realize fast rendering across various scales through the proposed block-wise detail levels selection and aggregation strategy. Extensive experimental results on large-scale scenes demonstrate that our approach attains state-of-theart rendering quality, enabling consistent real-time rendering of largescale scenes across vastly different scales. Our project page is available at
PDF Project Page:


Mirror-3DGS: Incorporating Mirror Reflections into 3D Gaussian Splatting

Authors:Jiarui Meng, Haijie Li, Yanmin Wu, Qiankun Gao, Shuzhou Yang, Jian Zhang, Siwei Ma

3D Gaussian Splatting (3DGS) has marked a significant breakthrough in the realm of 3D scene reconstruction and novel view synthesis. However, 3DGS, much like its predecessor Neural Radiance Fields (NeRF), struggles to accurately model physical reflections, particularly in mirrors that are ubiquitous in real-world scenes. This oversight mistakenly perceives reflections as separate entities that physically exist, resulting in inaccurate reconstructions and inconsistent reflective properties across varied viewpoints. To address this pivotal challenge, we introduce Mirror-3DGS, an innovative rendering framework devised to master the intricacies of mirror geometries and reflections, paving the way for the generation of realistically depicted mirror reflections. By ingeniously incorporating mirror attributes into the 3DGS and leveraging the principle of plane mirror imaging, Mirror-3DGS crafts a mirrored viewpoint to observe from behind the mirror, enriching the realism of scene renderings. Extensive assessments, spanning both synthetic and real-world scenes, showcase our method’s ability to render novel views with enhanced fidelity in real-time, surpassing the state-of-the-art Mirror-NeRF specifically within the challenging mirror regions. Our code will be made publicly available for reproducible research.
PDF 22 pages, 7 figures


Feature Splatting: Language-Driven Physics-Based Scene Synthesis and Editing

Authors:Ri-Zhao Qiu, Ge Yang, Weijia Zeng, Xiaolong Wang

Scene representations using 3D Gaussian primitives have produced excellent results in modeling the appearance of static and dynamic 3D scenes. Many graphics applications, however, demand the ability to manipulate both the appearance and the physical properties of objects. We introduce Feature Splatting, an approach that unifies physics-based dynamic scene synthesis with rich semantics from vision language foundation models that are grounded by natural language. Our first contribution is a way to distill high-quality, object-centric vision-language features into 3D Gaussians, that enables semi-automatic scene decomposition using text queries. Our second contribution is a way to synthesize physics-based dynamics from an otherwise static scene using a particle-based simulator, in which material properties are assigned automatically via text queries. We ablate key techniques used in this pipeline, to illustrate the challenge and opportunities in using feature-carrying 3D Gaussians as a unified format for appearance, geometry, material properties and semantics grounded on natural language. Project website:
PDF Project website:


StructLDM: Structured Latent Diffusion for 3D Human Generation

Authors:Tao Hu, Fangzhou Hong, Ziwei Liu

Recent 3D human generative models have achieved remarkable progress by learning 3D-aware GANs from 2D images. However, existing 3D human generative methods model humans in a compact 1D latent space, ignoring the articulated structure and semantics of human body topology. In this paper, we explore more expressive and higher-dimensional latent space for 3D human modeling and propose StructLDM, a diffusion-based unconditional 3D human generative model, which is learned from 2D images. StructLDM solves the challenges imposed due to the high-dimensional growth of latent space with three key designs: 1) A semantic structured latent space defined on the dense surface manifold of a statistical human body template. 2) A structured 3D-aware auto-decoder that factorizes the global latent space into several semantic body parts parameterized by a set of conditional structured local NeRFs anchored to the body template, which embeds the properties learned from the 2D training data and can be decoded to render view-consistent humans under different poses and clothing styles. 3) A structured latent diffusion model for generative human appearance sampling. Extensive experiments validate StructLDM’s state-of-the-art generation performance and illustrate the expressiveness of the structured latent space over the well-adopted 1D latent space. Notably, StructLDM enables different levels of controllable 3D human generation and editing, including pose/view/shape control, and high-level tasks including compositional generations, part-aware clothing editing, 3D virtual try-on, etc. Our project page is at:
PDF Project page:


MagicMirror: Fast and High-Quality Avatar Generation with a Constrained Search Space

Authors:Armand Comas-Massagué, Di Qiu, Menglei Chai, Marcel Bühler, Amit Raj, Ruiqi Gao, Qiangeng Xu, Mark Matthews, Paulo Gotardo, Octavia Camps, Sergio Orts-Escolano, Thabo Beeler

We introduce a novel framework for 3D human avatar generation and personalization, leveraging text prompts to enhance user engagement and customization. Central to our approach are key innovations aimed at overcoming the challenges in photo-realistic avatar synthesis. Firstly, we utilize a conditional Neural Radiance Fields (NeRF) model, trained on a large-scale unannotated multi-view dataset, to create a versatile initial solution space that accelerates and diversifies avatar generation. Secondly, we develop a geometric prior, leveraging the capabilities of Text-to-Image Diffusion Models, to ensure superior view invariance and enable direct optimization of avatar geometry. These foundational ideas are complemented by our optimization pipeline built on Variational Score Distillation (VSD), which mitigates texture loss and over-saturation issues. As supported by our extensive experiments, these strategies collectively enable the creation of custom avatars with unparalleled visual quality and better adherence to input text prompts. You can find more results and videos in our website:


NeRF-MAE : Masked AutoEncoders for Self Supervised 3D representation Learning for Neural Radiance Fields

Authors:Muhammad Zubair Irshad, Sergey Zakahrov, Vitor Guizilini, Adrien Gaidon, Zsolt Kira, Rares Ambrus

Neural fields excel in computer vision and robotics due to their ability to understand the 3D visual world such as inferring semantics, geometry, and dynamics. Given the capabilities of neural fields in densely representing a 3D scene from 2D images, we ask the question: Can we scale their self-supervised pretraining, specifically using masked autoencoders, to generate effective 3D representations from posed RGB images. Owing to the astounding success of extending transformers to novel data modalities, we employ standard 3D Vision Transformers to suit the unique formulation of NeRFs. We leverage NeRF’s volumetric grid as a dense input to the transformer, contrasting it with other 3D representations such as pointclouds where the information density can be uneven, and the representation is irregular. Due to the difficulty of applying masked autoencoders to an implicit representation, such as NeRF, we opt for extracting an explicit representation that canonicalizes scenes across domains by employing the camera trajectory for sampling. Our goal is made possible by masking random patches from NeRF’s radiance and density grid and employing a standard 3D Swin Transformer to reconstruct the masked patches. In doing so, the model can learn the semantic and spatial structure of complete scenes. We pretrain this representation at scale on our proposed curated posed-RGB data, totaling over 1.6 million images. Once pretrained, the encoder is used for effective 3D transfer learning. Our novel self-supervised pretraining for NeRFs, NeRF-MAE, scales remarkably well and improves performance on various challenging 3D tasks. Utilizing unlabeled posed 2D data for pretraining, NeRF-MAE significantly outperforms self-supervised 3D pretraining and NeRF scene understanding baselines on Front3D and ScanNet datasets with an absolute performance improvement of over 20% AP50 and 8% AP25 for 3D object detection.
PDF 29 pages, 13 figures. Project Page:


Surface Reconstruction from Gaussian Splatting via Novel Stereo Views

Authors:Yaniv Wolf, Amit Bracha, Ron Kimmel

The Gaussian splatting for radiance field rendering method has recently emerged as an efficient approach for accurate scene representation. It optimizes the location, size, color, and shape of a cloud of 3D Gaussian elements to visually match, after projection, or splatting, a set of given images taken from various viewing directions. And yet, despite the proximity of Gaussian elements to the shape boundaries, direct surface reconstruction of objects in the scene is a challenge. We propose a novel approach for surface reconstruction from Gaussian splatting models. Rather than relying on the Gaussian elements’ locations as a prior for surface reconstruction, we leverage the superior novel-view synthesis capabilities of 3DGS. To that end, we use the Gaussian splatting model to render pairs of stereo-calibrated novel views from which we extract depth profiles using a stereo matching method. We then combine the extracted RGB-D images into a geometrically consistent surface. The resulting reconstruction is more accurate and shows finer details when compared to other methods for surface reconstruction from Gaussian splatting models, while requiring significantly less compute time compared to other surface reconstruction methods. We performed extensive testing of the proposed method on in-the-wild scenes, taken by a smartphone, showcasing its superior reconstruction abilities. Additionally, we tested the proposed method on the Tanks and Temples benchmark, and it has surpassed the current leading method for surface reconstruction from Gaussian splatting models. Project page:
PDF Project Page:


文章作者: 木子已
版权声明: 本博客所有文章除特別声明外,均采用 CC BY 4.0 许可协议。转载请注明来源 木子已 !