This work proposes a visual relocalization method that departs from classical correspondence-based pipelines by directly estimating camera poses against a differentiable map representation built with 3D Gaussian Splatting (3DGS), and shows substantial gains in relocalization accuracy under challenging conditions.
Abstract
Visual localization becomes extremely challenging in planetary-like terrains characterized by low texture, perceptual aliasing, harsh illumination, and sparse, weakly overlapping viewpoints induced by forward rover motion and unconstrained driving directions. Under these conditions, state-of-the-art image-to-image and image-to-map matching pipelines suffer significant performance degradation. In this work, we propose a visual relocalization method that departs from classical correspondence-based pipelines by directly estimating camera poses against a differentiable map representation built with 3D Gaussian Splatting (3DGS). Our key contribution is a geometry-aware training strategy that combines photometric and geometric losses, where the geometric supervision is provided for the first time by combining multi-view stereo (MVS) and LiDAR depths. We show that this joint optimization produces a 3DGS model that better fits the underlying scene geometry, leading to improved photometric and geometric consistency and more robust, accurate single-image 6-DoF pose estimation. Extensive experiments on data acquired in planetary-analog environments validate the effectiveness of our approach, showing substantial gains in relocalization accuracy under challenging conditions. Code is available at https://github.com/DLR-RM/multimodal-gsplat-relocalization.
This work enhances the existing iterative object-basesd visual localization approach with an additional semantic feature derived from a pretrained semantic segmentation model and conducts a systematic baseline study of contemporary feature matching techniques on such cross-domain query-reference image pairs.
Yasmin Loeper, Markus Gerke, P. Fanta-Jende· The International Archives o...· 0 citations
The task of reconstructing dynamic neural scenes from sparse multi-view observations represents a significant challenge in computer vision and computer graphics. Traditional neural rendering techniques require dense view sampling to synthesize high-quality novel views, which is highly impractical for real-world dynamic environments where deploying numerous synchronized cameras is prohibitively expensive and logistically complex. When constrained to sparse views, existing dynamic reconstruction frameworks typically suffer from severe overfitting, resulting in pronounced geometric distortions, floaters, and temporal inconsistencies. This paper proposes a novel framework designed to achieve high-fidelity dynamic neural scene reconstruction from highly sparse viewpoints by integrating geometry-aware depth priors and robust multi-view correspondence constraints. By leveraging monocular depth estimation aligned with sparse structural cues, the system enforces a strict geometric foundation that prevents the neural field from degenerating in unobserved regions. Furthermore, we introduce a cross-view feature correspondence mechanism that penalizes photometric and geometric divergences across temporal and spatial domains, ensuring consistency in the deformation fields used to model scene dynamics. Comprehensive evaluations demonstrate that our methodology significantly suppresses artifacts and achieves superior novel view synthesis quality compared to baseline methods. The integration of these complementary constraints effectively bridges the information gap inherent in sparse observations, offering a scalable solution for dynamic scene capture.
Camila Wilson, D. J. Ross· Journal of innovative resear...· 0 citations
Reconstructing high-fidelity 3D scenes from sparse-views remains a central problem in generalizable neural rendering. Existing generalizable 3D Gaussian Splatting (3DGS) methods often exhibit geometric artifacts in sparse-view settings, since supervision based solely on 2D photometric losses cannot resolve depth and correspondence ambiguities. To address this issue, we propose MAC-Splat, a training framework built around direct 3D consistency supervision. MAC-Splat builds on the MASt3R geometric backbone and a frozen DINOv3 encoder to obtain semantically informed 2D correspondences, which serve as geometric anchors for 3D supervision. Using these anchors, we define the Multi-Attribute Consistency (MAC) loss. This objective jointly regularizes the 3D attributes of matched Gaussians, including their position, shape, and appearance, by enforcing agreement in a common world coordinate frame. The formulation is robust to outliers and respects the geometry of covariance matrices, which leads to stable training under sparse-view conditions. Experiments on ScanNet++ show that MAC-Splat outperforms strong baselines, with particularly large gains under different overlap regimes. In particular, it improves average PSNR over Splatt3R by more than 4.5 dB, reduces LPIPS, and maintains performance as the camera pose gap increases. These results indicate that a direct, multi-attribute 3D consistency objective, when combined with high-quality correspondences, is effective for addressing the ill-posed sparse-view reconstruction problem.
Jinqian Yang, Yichen Wu, Wanhua Li et al.· 1 citation
A semantic-guided 3D Gaussian splatting (3DGS) framework tailored to sparse-view industrial reconstruction was introduced, enabling robust reconstruction from limited viewpoints and offers a practical geometric foundation for automated inspection and remote equipment monitoring.
Boyang Li, Tianhan Gao, Zuan Gu et al.· Visual Computing for Industr...· 0 citations
3D Gaussian splatting (3DGS) provides an efficient and expressive scene representation by jointly modeling spatial geometry and appearance, which has led to significant advances in high-fidelity scene reconstruction and novel view synthesis. However, in real-world environments, occlusions from pedestrians or equipment often introduce blurriness, artifacts, and geometric distortions. To address these challenges, this paper proposes a robust 3DGS modeling method enhanced by semantic and geometric consistency. First, the self-supervised foundation model DINOv2 is utilized to extract high-dimensional semantic features, leveraging its superior generalization capabilities to assist in identifying potential dynamic regions. Second, monocular depth estimation and a depth residual mechanism are introduced to construct geometric consistency constraints, enabling the precise localization of areas that violate static assumptions. Finally, a progressive guided probability masking mechanism is designed; it employs an adaptive sigmoid function to achieve a ‘coarse-to-fine’ soft-constraint optimization, effectively mitigating the training instability inherent in traditional binary hard masks. Experimental results on the neural radiance fields (NeRF)-on-the-go, RobustNeRF, and self-collected datasets demonstrate that the proposed method effectively suppresses dynamic artifacts and improves reconstruction quality. The proposed approach achieves competitive or superior performance compared with 3DGS, SpotLessSplats, T-3DGS, and RobustSplat on standard image-quality metrics, including peak signal-to-noise ratio, structural similarity index measure, and learned perceptual image patch similarity.
Wen Zheng, Guo Bao, Wenda Wang et al.· Engineering Research Express· 0 citations
Classical image correspondence is solved at the level of sparse keypoints or dense pixels, but the systems that consume these matches - object-level mapping, topological navigation, scene-graph maintenance - reason about whole objects. Recent work narrows this gap by matchng directly at the level of instance segments: a class-agnostic segmenter partitions each image, and per-segment descriptors are obtained by pooling features from large 3D foundation models over the masks. We build on this segment-level matching paradigm and propose three learned matching heads: a LightGlue-style attention head with DoubleSoftmax scoring on frozen MASt3R descriptors; a DPT-style multi-scale fusion module that exposes layered spatial detail from the VGGT foundation model before pooling; and - as our main contribution - a multi-view extension that performs joint self-attention over segments drawn from several views at once, recovering transitive correspondences that strictly pairwise matchers cannot reach. Under a stratified zero-shot protocol on Replica and Virtual KITTI 2 with controlled viewpoint baselines from 0 deg to 180 deg, the LightGlue-style head improves over a parameter-free Sinkhorn matcher on the same MASt3R backbone by +4.85 AUPRC on Replica and +25.9 AUPRC on Virtual KITTI 2. Dropped into the RoboHop topological navigation pipeline on the Habitat-Matterport 3D (HM3D) Instance Image Navigation benchmark without retraining, our multi-view variant raises success rate from 50% to 70%, and our LightGlue-style head raises SPL from 45.7 to 59.1.
Denis Fatykhoph, Timur Akhtyamov, Konstantin Pakulev et al.· 0 citations