gravity-aligned 3d reconstruction

This work addresses the limitation of existing RGB-based multi-view reconstruction methods, which produce monolithic scene representations lacking explicit physical structure and thus hinder stable physical interaction. The authors propose an end-to-end reconstruction framework that relies solely on RGB images and introduces gravity as a universal physical prior. By aligning views in a gravity-consistent coordinate system, reconstructing object-centric rigid-body meshes, and employing conditional 3D point classification to remove background redundancy, the method decouples foreground objects from background geometry without requiring CAD model retrieval. The output is a structured hybrid representation suitable for simulation. Experiments demonstrate significant improvements over retrieval-based baselines in 6-DoF object pose accuracy, decoupling quality, and rendering-to-simulation efficiency, both in simulated and real-world scenes.

This work addresses the limitations of traditional 3D reconstruction methods, which predict point maps in camera-centered coordinates, struggle to incorporate scene structural priors, and suffer from high rotational degrees of freedom across views, leading to inconsistent reconstructions. To overcome these issues, the authors propose predicting point maps in a gravity-aligned upright coordinate system, thereby reducing inter-view rotational ambiguity through a shared vertical axis. They introduce the Gravity Grounded Geometry Transformer (G3T) model and the G3T-Long incremental reconstruction framework, which for the first time integrate gravity-aligned coordinates into point map prediction by combining a Transformer architecture, gravity-aware pose estimation, and a submap stitching strategy. Experiments demonstrate that this approach significantly improves reconstruction accuracy and robustness, outperforming existing methods in incremental 3D reconstruction and validating the effectiveness of gravity-aligned representations.

This work addresses the challenge of accurate image-based localization in real-world complex environments, where existing methods rely on small-scale vectorized floorplans and struggle with ordinary images. The authors propose a novel approach that first reconstructs a gravity-aligned 3D scene from unconstrained input images to generate a 2D density map as a proxy for the floorplan, then aligns this map with the given floorplan via a 2D similarity transformation. By innovatively integrating 3D reconstruction with 2D foundation models, the method employs a fine-tuning strategy that enforces semantic alignment and structural consistency, enabling robust performance even with extremely sparse inputs—such as a single image. Experiments demonstrate that the proposed technique significantly outperforms state-of-the-art methods across diverse real-world scenarios, achieving reliable localization with minimal data.

This work addresses the lack of physical plausibility in single-image 3D reconstruction. We propose the first physics-compatible reconstruction framework that enforces static equilibrium as a hard constraint. Methodologically, we explicitly decouple and jointly optimize material stiffness, external loading forces, and the static equilibrium geometry; deformation responses are modeled via differentiable physics simulation, enabling gradient-based joint optimization of all variables. Our approach breaks from conventional simplifications—such as rigid-body assumptions or neglect of external forces—by embedding real-world physical constraints directly into the single-image reconstruction pipeline. Evaluated on Objaverse, our method yields reconstructions with significantly improved mechanical stability, suitable for downstream dynamic simulation and 3D printing. Physical validation via real-world force testing further confirms the structural robustness of the generated models.

In large-scale Structure-from-Motion (SfM), sparse inter-view overlap and drastic viewpoint changes—especially in aerial-to-ground scenarios—lead to low cross-image feature matching density and weak geometric consistency. To address this, we propose a geometry-guided hybrid matching paradigm: (1) geometric verification is formulated as an optimization problem based on Sampson distance; (2) detector-agnostic dense matching is fused with detector-driven sparse anchor guidance, where sparse anchors constrain and enhance the geometric consistency of dense matches; and (3) multi-view geometric consistency is explicitly modeled. Our method significantly improves both matching density and accuracy, outperforming state-of-the-art approaches in extreme large-scale settings. Consequently, camera pose estimation becomes more accurate, and the reconstructed 3D point cloud achieves higher completeness and fidelity.

This study addresses the spatial distortion in oblique-view remote sensing imagery caused by geometric projection displacement between building rooftops and their ground footprints. To this end, it formulates roof-to-footprint offset vector (RFOV) extraction as an independent learning task, decoupling geometric correction from semantic segmentation. The authors introduce ObliCity, the first large-scale oblique urban dataset, which integrates high-resolution drone and satellite imagery, and propose DragRoof—a framework based on ordinary differential equations (ODEs) that simulates a continuous, annotation-inspired dragging process to adaptively learn geometrically consistent offset fields. Experiments demonstrate that the proposed method achieves state-of-the-art RFOV extraction accuracy on ObliCity with fewer inference steps, significantly outperforming existing approaches in both direction and magnitude estimation.

This study addresses the significant challenge of reconstructing watertight, 3D-printable models from rover imagery in Martian terrain, which is characterized by low texture, irregular geometry, and incomplete observations. The authors propose and systematically evaluate the first end-to-end pipeline tailored for Mars scenes, integrating RAFT-Stereo and SGBM for stereo matching, followed by geometric completion using Alpha Shapes, Poisson surface reconstruction, and deterministic diffusion-based hole filling to produce watertight meshes. Experimental results reveal that RAFT-Stereo—despite its strong performance on terrestrial datasets—exhibits inferior edge alignment on Martian images, and that the effectiveness of geometric completion methods varies substantially, underscoring the necessity of domain-specific validation for Mars missions. This work establishes the first benchmark demonstrating the limitations of general-purpose 3D reconstruction techniques in extraterrestrial environments and provides critical guidance for manufacturable modeling in planetary exploration.

GALoc通过使用重力对齐的线框图替代深度预测，解决了室内定位中深度网络在杂乱场景中的脆弱性问题。

Existing single-view mesh reconstruction methods exhibit poor generalization under camera rotation due to their reliance on viewpoint priors, often resulting in 3D inconsistencies, erroneous scene layouts, and violations of physical constraints. This work proposes the first evaluation protocol specifically designed for single-view reconstruction under camera rotation, enabling systematic assessment of depth estimation, object meshes, scene layout, and physical plausibility. A two-stage pipeline built upon SAM3D and FoundationPose—augmented with ICP registration, monocular depth estimation, and gravity alignment—significantly enhances robustness. Furthermore, a novel gravity-aware refinement strategy reduces layout orientation error by 47.1% compared to single-stage approaches.

High-fidelity 3D surface reconstruction of large-scale urban scenes remains challenging due to geometric complexity, prolonged optimization times, and GPU memory constraints. This work proposes a viewpoint-direction-based scene partitioning strategy that groups views with similar orientations for joint optimization, substantially improving depth estimation accuracy and enabling balanced multi-GPU computation. Additionally, a point cloud hole detection and inpainting mechanism is introduced to enhance geometric completeness. Built upon the 3D Gaussian Splatting (3DGS) framework, the method integrates partitioned optimization, parallel computation, and scene completion techniques. Extensive experiments on GauU-Scene, MatrixCity, and UrbanScene3D datasets demonstrate that the proposed approach significantly outperforms state-of-the-art methods in reconstruction quality.