# SEAR: Simple and Efficient Adaptation of Visual Geometric Transformers for Unpaired RGB+Thermal 3D Reconstruction Source: [SEAR: Simple and Efficient Adaptation of Visual Geometric Transformers for Unpaired RGB+Thermal 3D Reconstruction]() Collection index: [llms.txt]() Data updated: 2026-08-07T05:00:01.339Z Authors: Vsevolod Skorokhodov, Chenghao Xu, Shuo Sun, Olga Fink, Malcolm Mielle Published: 2026-03-19T11:27:21.000Z ## Abstract Foundational feed-forward visual geometry models enable accurate and efficient camera pose estimation and scene reconstruction by learning strong scene priors from massive RGB datasets. However, their effectiveness drops when applied to mixed sensing modalities, such as RGB-thermal (RGB-T) images. We observe that while a visual geometry grounded transformer pretrained on RGB data generalizes well to thermal-only reconstruction, it struggles to align RGB and thermal modalities when processed jointly. To address this, we propose SEAR, a simple yet efficient fine-tuning strategy that adapts a pretrained geometry transformer to multimodal RGB-T inputs. Despite being trained on a relatively small RGB-T dataset, our approach significantly outperforms state-of-the-art methods for 3D reconstruction and camera pose estimation, achieving significant improvements over all metrics and delivering higher detail and consistency between modalities with negligible overhead in inference time compared to the original pretrained model. Notably, SEAR enables reliable multimodal pose estimation and reconstruction even under challenging conditions, such as low lighting and dense smoke. We validate our architecture through extensive ablation studies and demonstrate how the model aligns both modalities. Additionally, we introduce a new dataset featuring RGB and thermal sequences captured at different times, viewpoints, and illumination conditions, providing a robust benchmark for future work in multimodal 3D scene reconstruction. Code and models are publicly available at https://doi.org/10.5281/ZENODO.21077295. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [3D reconstruction on Public Datasets](): 3D Reconstruction; Public datasets; AUC: 70. - [3D reconstruction on SEAR Dataset](): 3D Reconstruction; SEAR Dataset; AUC: 62.8. - [Camera Pose Estimation on ViViD++](): Camera pose estimation; ViViD++; RRA@30: 91. - [Camera Pose Estimation on STheReO](): Camera pose estimation; STheReO; RRA@30: 100. - [Depth Estimation on ViViD++ (indoor\_aggresive\_dark, indoor\_unstable\_dark, outdoor\_robust\_night2)](): Depth Estimation; ViViD++ indoor\_aggresive\_dark, indoor\_unstable\_dark, outdoor\_robust\_night2; AbsRel: 0.1956. - [Depth Estimation on STheReO kaist\_evening, valley\_evening](): Depth Estimation; STheReO kaist\_evening, valley\_evening; AbsRel: 0.237. - [Depth Estimation on Dark3R (dark\_chapel, dark\_kitchen)](): Depth Estimation; Dark3R (dark\_chapel, dark\_kitchen); Abs Rel: 0.1703. - [Camera Pose Estimation on Dark3R](): Camera pose estimation; Dark3R; RRA (30 deg): 66. ## Original sources and implementation [Original paper]() [PDF]()