This work introduces a system that combines the strong sequential constraints of SLAM with the flexibility and global optimization of offline SfM, enabling the metric reconstruction of arbitrary, long, uncalibrated videos.
Abstract
Accurately recovering the camera's calibration and metric poses for any unconstrained video would unlock large-scale training data for navigation and scene understanding. The dominant approaches to this problem are severely limited: Simultaneous Localization and Mapping (SLAM) is sensitive to initialization and transient failures due to its causal, incremental nature; it is often over-optimized for real-time operation and generally requires known camera calibration; while Structure-from-Motion (SfM) typically forgoes any image ordering, enabling optimal initialization and global optimization, but lacks robustness to visual symmetries and extreme motions. To bridge this gap, we introduce a system that combines the strong sequential constraints of SLAM with the flexibility and global optimization of offline SfM, enabling the metric reconstruction of arbitrary, long, uncalibrated videos. This system leverages recent advances in wide-baseline dense image matching, treats temporal ordering as a first-class citizen for reliable loop closure, and augments global optimization with metric monocular depth priors. As a result, thorough evaluations on diverse, challenging datasets that exhibit extreme motion and visual symmetries reveal that our approach is significantly more robust and accurate than both state-of-the-art SLAM and SfM, classical or learned, with given or unknown camera calibration. The code is publicly available at https://github.com/cvg/vidmap.
Global Structure-from-Motion (SfM) is an efficient paradigm for recovering camera poses and sparse 3D structure from unordered images. However, its reliance on scale-ambiguous epipolar geometry makes global positioning sensitive to noisy baseline estimates and weak view-graph constraints, while false edges from visually ambiguous pairs can further degrade reconstruction. We propose DGSfM, a depth-aware global SfM pipeline that uses monocular depth maps as a scalable prior while preserving explicit multi-view optimization. For each image pair, we use a depth-aware relative pose solver to convert scale-ambiguous epipolar constraints into scale-aware relative pose constraints. We further improve robustness through view-graph filtering and depth-consistency-based correspondence pruning, which suppress false edges and matches that remain plausible under epipolar geometry alone. Finally, global scale averaging and depth-guided pose-point initialization align monocular depth maps into a common reconstruction scale and provide stable initialization for global positioning and bundle adjustment. Experiments on ETH3D and IMC2021 show that DGSfM consistently improves over strong global SfM baselines across sparse and dense matching front-ends, achieving substantial gains in pose accuracy. Code is available at https://github.com/sithu31296/DGSfM.
Sithu Aung, Viktor Kocur, Yaqing Ding et al.· 0 citations
Monocular RGB dense SLAM remains challenging because depth and global metric scale are not directly observable from a single camera. Existing systems often combine optical-flow and monocular-geometry priors predicted by independently trained networks, which can provide inconsistent constraints to bundle adjustment (BA). Our quantitative prior-consistency analysis indicates that this disagreement is an important contributor to unstable local optimization and reconstruction error rather than the sole cause of drift. We propose KP-SLAM, which predicts dense optical flow and paired pointmap priors from a shared representation and incorporates them into the same BA backend. We further introduce a Depth-Scale-Pose-to-Pointmap (DSPP) objective that relates optimized inverse depth, edge-wise relative scale, and camera pose to paired pointmap constraints. Experiments on ScanNet, TUM-RGBD, KITTI, Tanks-and-Temples, and dynamic sequences show improved tracking, depth, and rendering metrics over the compared RGB-only baselines under the reported settings. The results support the usefulness of synchronized priors while also revealing remaining limitations in highly dynamic, weakly textured, and large-scale scenes.
Song Gao, Xinyu Huang, Zheng Huang et al.· Symmetry· 0 citations
Recent 3D geometric foundation models, such as VGGT, provide robust feed-forward 3D reconstruction by directly predicting camera poses and 3D scene points from input images. However, their results remain inaccurate, and scaling them to long sequences or large unordered image sets typically requires chunk-wise processing, which can introduce drift and inconsistency. We present Glob3R, a global SfM-style reconstruction built on 3D foundation models. Our key idea is to explicitly optimize feed-forward geometric predictions. To this end, we augment a frozen Pi3X backbone with a lightweight dense matching head that predicts image warps between selected reference frames and neighboring views. These dense warps are converted into sparse but reliable multi-view feature tracks, which provide correspondence constraints for global optimization. We further introduce a keyframe-based sliding-window association strategy that propagates tracks and relative poses across overlapping windows, enabling scalable reconstruction. Finally, we perform global motion averaging and bundle adjustment to refine camera poses, reduce scale inconsistencies, and recover dense scene geometry. Extensive experiments on indoor, outdoor, large-scale driving, and unordered SfM benchmarks demonstrate that Glob3R achieves robust and accurate reconstruction. It consistently improves over feed-forward foundation-model baselines and recent scalable reconstruction methods, while being more robust than classical SfM pipelines. The refined poses also lead to higher-quality neural rendering, validating the benefit of combining foundation-model priors with global geometric optimization. Project page: https://junyuandeng.github.io/Glob3r
Junyuan Deng, Heng Li, Ke-Jie Qiu et al.· 0 citations
Motion-only multiobject tracking (MOT) suffers from ID switches in uniform-appearance and deformation-heavy scenes. In these settings, appearance cues become less reliable, so stable identities depend mainly on motion information. Existing methods often process all bounding-box variables together, which can weaken cues needed for prediction and matching. We address this problem by treating each trajectory as a heterogeneous multivariate time series (MTS) and redesigning the motion-only pipeline for embedding, encoding, and matching. HSMTrack separates box-state variables before modeling their temporal and cross-variable relationships, then uses deformation-aware matching for identity association. The method requires no appearance branch and can serve as a post-detection motion module in vision-sensor tracking pipelines. Its SSM-based encoder has linear complexity with respect to trajectory length, reducing modeling cost compared with attention-based alternatives. HSMTrack achieves 59.6 IDF1 and 42.9 AssA on DanceTrack, and 77.9 IDF1 and 67.2 AssA on SportsMOT. Under a unified end-to-end protocol, it reaches 34.1 frames/s on RTX 4090 and 10.5 frames/s on Jetson Orin NX.
Tianjing Cheng, Qingyuan Yu, Bo Jiang et al.· IEEE Sensors Journal· 0 citations
Structure-from-Motion (SfM) is a cornerstone of 3D perception, yet current methods often fail when applied to complex videos involving challenging camera motions or dynamic scenes. Compounding the problem, the field lacks reliable ground-truth benchmarks for such difficult scenarios, making it hard to gauge real-world progress or to pinpoint where improvements are most needed. To address this gap, we introduce a new benchmark for evaluating camera pose estimation. Our key insight is to leverage online panoramic 360{\deg} video as a source of data from which to construct challenging clips, while still enabling robust ground-truth trajectory recovery. The panoramic nature of these videos provides richer visual context for tracking camera motion, even when parts of the view are affected by blur, motion, or dynamic objects. After tracking camera motion across full 360{\deg} videos, we crop and reproject selected portions to generate perspective-view clips that serve as our benchmark, called ORBIT. Experiments show that COLMAP, as well as recent optimization-based and feed-forward SfM methods struggle to accurately estimate camera poses on our benchmark. Hence, ORBIT provides a valuable testbed where researchers can meaningfully measure progress on truly challenging, real-world SfM problems.
S. Sabour, Linyi Jin, Richard Tucker et al.· 0 citations
Image-goal navigation is a fundamental capability for embodied AI, yet its practical deployment is strained by strong prior assumptions. Existing methods predominantly rely on temporally ordered video streams or auxiliary sensors (e.g., depth, LiDAR) to maintain spatial consistency. These sequential and multimodal dependencies severely restrict scalability, especially when deploying robots using crowd-sourced or pre-recorded unordered image collections. When temporal priors are removed, current methods struggle with severe perceptual aliasing, noisy associations, and catastrophic mapping failures. To address this underexplored challenge, we propose Unordered Landmark Visual Navigation (ULVN), a unified RGB-only framework free from temporal and odometric priors. ULVN systematically mitigates error accumulation by integrating mapping, localization, and planning. Specifically, it constructs a robust 2D topological map directly from unstructured images via calibrated geometric verification and maximum spanning forest refinement. For closed-loop execution, ULVN abandons sequential heuristics, utilizing a graph-based belief propagation filter with entropy-adaptive fusion for global localization and dynamic subgoal planning. Extensive experiments in simulation and real-world deployments demonstrate that ULVN significantly outperforms state-of-the-art methods.