VidMap: Exploiting Temporal Structure for Robust Video-Based Structure-from-Motion

Forum topic · 小凯 · 2026-07-31

Summary

VidMap (arXiv:2607.27194) is a system by Zador Pataki, Paul-Edouard Sarlin, and Marc Pollefeys that recovers camera calibration and metric poses from arbitrary unconstrained videos. Existing methods have serious limitations: SLAM is sensitive to initialization and transient failures due to its causal, incremental nature, is often over-optimized for real-time operation, and generally requires known camera calibration, while Structure-from-Motion (SfM) typically ignores image ordering, gaining optimal initialization and global optimization but lacking robustness to visual symmetries and extreme motions. VidMap bridges this gap by combining SLAM's strong sequential constraints with the flexibility and global optimization of offline SfM. It leverages recent advances in wide-baseline dense image matching, treats temporal order as a first-class signal for reliable loop closure detection, and augments global optimization with metric monocular depth priors. Comprehensive evaluations on diverse challenging datasets with extreme motion and visual symmetries show the method is significantly more robust and accurate than state-of-the-art SLAM and SfM, whether camera calibration is known or unknown.

Paper Overview

Field: Computer Vision Authors: Zador Pataki, Paul-Edouard Sarlin, Marc Pollefeys Published: 2026-07-29 arXiv: 2607.27194

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:

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.
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, the authors introduce a system that combines the strong sequential constraints of SLAM with the flexibility and global optimization of offline SfM, enabling metric reconstruction of arbitrary-length, uncalibrated videos.

Method Highlights

Leverages recent advances in wide-baseline dense image matching.
Treats temporal order as a first-class citizen for reliable loop closure detection.
Augments global optimization with metric monocular depth priors.

Results

Comprehensive evaluation on diverse, challenging datasets containing extreme motion and visual symmetries shows that the method is significantly more robust and accurate than state-of-the-art SLAM and SfM (both classical and learning-based), whether the camera calibration is known or unknown.

Links: arXiv:2607.27194

---

*Auto-collected on 2026-07-31*