跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.14763 · v1

RECALL: Real-Time Recalibration of Multi-Camera Rigs via Learned OEM-Anchored Residuals

首次提交 2026-11-11 05:39
版本更新 2026-11-11 05:39
雷达观测 2026-11-13 10:09
arXiv ↗PDF ↗

摘要

Multi-camera rigs ship pre-calibrated from the manufacturer, but their calibration goes stale each time the rig is repositioned, and drifts continuously through wear, thermal cycling, and vibration. Recovering it on-the-fly is hard: classical bundle adjustment needs a fresh multi-view capture and a global optimization whose cost grows with the rig, ruling out real-time use. In this paper we present RECALL, a feed-forward method that recalibrates a movable multi-camera rig in a single pass. From the nominal OEM calibration as an anchor and a brief observation of a known object, it predicts each camera's current intrinsics and extrinsics, written as a shared global rotation of the rig composed with a small per-camera drift. Instead of regressing the calibration in one shot, RECALL refines its estimate with a learned, rig-coupled recurrent operator that reprojects through a differentiable projection and applies attention across the cameras at each step, and solves for the shared rotational gauge in-network with a learned Gauss-Newton step rather than regressing it. Because the ground truth is synthesizable, we train entirely on data generated on-the-fly and need no captured dataset. On a synthetic 10-camera benchmark, RECALL is more accurate than per-camera and rig-coupled joint bundle adjustment (0.26∘0.26^\circ vs 0.47∘0.47^\circ rotation, 5.45.4 vs 8.98.9 px focal) and runs about 105×10^5\times faster at about 0.120.12 ms per rig. On real data (TotalCapture, CMU Panoptic) it matches or beats offline rig-coupled bundle adjustment on absolute and relative rotation while running four-to-five orders of magnitude faster, and on the Panoptic surround rig it also matches the strongest image-only foundation models (VGGT, VGGT-ΩΩ, π3π^3) on relative pose while staying metric and about 15×15\times more accurate in focal length. Code: https://github.com/theICTlab/RECALL-multicam

作者备注:WACV 2027 submission

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-11 05:39 · 最新观测版本

RECALL: Real-Time Recalibration of Multi-Camera Rigs via Learned OEM-Anchored Residuals

Multi-camera rigs ship pre-calibrated from the manufacturer, but their calibration goes stale each time the rig is repositioned, and drifts continuously through wear, thermal cycling, and vibration. Recovering it on-the-fly is hard: classical bundle adjustment needs a fresh multi-view capture and a global optimization whose cost grows with the rig, ruling out real-time use. In this paper we present RECALL, a feed-forward method that recalibrates a movable multi-camera rig in a single pass. From the nominal OEM calibration as an anchor and a brief observation of a known object, it predicts each camera's current intrinsics and extrinsics, written as a shared global rotation of the rig composed with a small per-camera drift. Instead of regressing the calibration in one shot, RECALL refines its estimate with a learned, rig-coupled recurrent operator that reprojects through a differentiable projection and applies attention across the cameras at each step, and solves for the shared rotational gauge in-network with a learned Gauss-Newton step rather than regressing it. Because the ground truth is synthesizable, we train entirely on data generated on-the-fly and need no captured dataset. On a synthetic 10-camera benchmark, RECALL is more accurate than per-camera and rig-coupled joint bundle adjustment (0.26∘0.26^\circ vs 0.47∘0.47^\circ rotation, 5.45.4 vs 8.98.9 px focal) and runs about 105×10^5\times faster at about 0.120.12 ms per rig. On real data (TotalCapture, CMU Panoptic) it matches or beats offline rig-coupled bundle adjustment on absolute and relative rotation while running four-to-five orders of magnitude faster, and on the Panoptic surround rig it also matches the strongest image-only foundation models (VGGT, VGGT-ΩΩ, π3π^3) on relative pose while staying metric and about 15×15\times more accurate in focal length. Code: https://github.com/theICTlab/RECALL-multicam

查看此版本解读 →查看此版本原文 ↗