cs.CVNov 11, 2026

ThermalSplat: Generalizable Pose-Free Gaussian Splatting for Sparse-View Thermal 3D Reconstruction

Authors: Prasanna S, Gopi Raju Matta, Divya Madhuri Vemunuri, Kaushik Mitra

Organizations: Indian Institute of Technology Madras

Abstract

Thermal infrared imaging is vital for applications like search-and-rescue and autonomous navigation, yet trans- lating it into 3D scene representations remains a critical bottleneck. The low-texture nature and view-dependent emissivity of thermal images cause standard pose estima- tors to fail catastrophically, collapsing standard 3D Gaus- sian Splatting pipelines under sparse-view constraints. Fur- thermore, existing thermal reconstruction methods rely heavily on dense capture and computationally expensive per-scene optimization. Critically, many of these ap- proaches depend on aligned RGB-thermal datasets to ob- tain reliable camera poses and initial point clouds, ren- dering them unusable in purely thermal, unposed envi- ronments. To break this bottleneck, we present Thermal- Splat, the first generalizable, feed-forward Gaussian splat- ting framework that reconstructs 3D scenes directly from sparse thermal images without requiring known camera poses, RGB guidance, or per-scene training. We over- come thermal-specific challenges using a soft census loss for brightness-invariant structural matching and a bound- ary structure loss to explicitly supervise temperature discon- tinuities. Trained on an aggregated multi-source dataset, ThermalSplat synthesizes sharp reconstructions of unseen scenes in a single feed-forward pass. Extensive experi- ments demonstrate that ThermalSplat holds a strict advan- tage over scene-specific thermal splatting methods in the sparse regime. Furthermore, even compared against these baselines operating under their ideal dense capture set- tings, ThermalSplat achieves superior perceptual fidelity. By eliminating the need for computationally expensive per scene optimization, ThermalSplat offers an efficient and generalizable approach to thermal 3D reconstruction.

Figures & tables

Appendix figures & tables7 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

Jun 26, 2026

StructSplat: Generalizable 3D Gaussian Splatting from Uncalibrated Sparse Views

We present StructSplat, a feed-forward and generalizable 3D Gaussian reconstruction framework that operates directly on uncalibrated images without requiring camera parameters. Existing methods either rely on per-scene optimization or assume known camera poses, and often entangle geometry and appearance within a unified backbone, limiting reconstruction fidelity and generalization. Our key idea is to adopt a structured representation that organizes geometry, semantic, and texture cues with explicit roles in the reconstruction process. Specifically, we introduce a pixel-aligned feature injection mechanism to enable accurate texture modeling from 2D observations, incorporate semantic-aware priors to improve global consistency, and design a camera alignment strategy to prevent information leakage and improve generalization. Experiments show that our method significantly outperforms prior approaches on challenging benchmarks. On DL3DV, our method achieves 28.045 PSNR, surpassing AnySplat (22.377) by +5.67 dB. In cross-dataset evaluation, our method achieves +1.94 dB over AnySplat on ACID and +1.72 dB on RealEstate10K. Project page: https://structsplat.github.io Code: https://github.com/J-C-Zhao/StructSplat
May 28, 2026

Supercharging Thermal Gaussian Splatting with Depth Estimation

Efficient and robust 3D scene representation is crucial in autonomous driving, robotics, and related fields. While RGB images provide valuable content for 3D reconstruction, other modalities like thermal or depth can enable additional information on the environment. Lately, novel view synthesis methods like 3D Gaussian Splatting have started using multiple modalities to further boost their performance. But fusing or combining multimodal data can make the process slower and can bring in additional challenges. Therefore, our project aims to use single modality based on thermal infrared domain, by removing the reliance on visible light as much as possible. This single modality can be expected to be faster as it does not rely on multimodal data. We propose a method, Thermal-to-Depth Gaussian Splatting (TDg), that uses only thermal images and depth estimation in its architecture to derive the radiance fields. Our TDg method outperforms the MSMG (Multiple Single-Modal Gaussians) baseline in most cases on our test datasets, RGBT-Scenes and ThermalMix. On average, the rendering quality metrics such as learned perceptual image patch similarity (LPIPS), structural similarity index measure (SSIM), and peak signal-to-noise ratio (PSNR) of TDg are 1.12%, 0.034%, and 0.01% better than the baseline MSMG values. It also reduces the training time significantly, by 12 mins 47 secs (55% improvement). Overall, our method is successful in deriving these thermal radiance fields, which can ultimately have several applications, such as identifying heat sources critical in surveillance, search or rescue operations, and industrial inspections where temperature is widely used to monitor machines.
Jun 3, 2026

Unpaired RGB-Thermal Gaussian-Splatting Using Visual Geometric Transformers

Multi-modal novel view synthesis (NVS) combining RGB and thermal imagery enables precise 3D scene reconstruction with visual and thermal information. However, existing methods typically rely on precisely calibrated RGB-thermal image pairs or stereo setups, limiting scalability and practical deployment. To address this, we introduce a framework for unpaired RGB-thermal NVS that leverages VGGT, a 3D feed-forward transformer architecture, to independently estimate camera poses for each modality. The pose sets are then aligned using the Procrustes algorithm with a cross-modal feature matcher, enabling joint registration without paired calibration. Building on this alignment, we further propose a multi-modal 3D Gaussian Splatting approach that learns directly from unpaired RGB and thermal images. Experiments on diverse scenes demonstrate that our method achieves competitive performance in thermal view synthesis while maintaining RGB fidelity. Moreover, we show that existing reconstruction approaches can produce modality-specific reconstructions that lack cross-modal consistency. We thus introduce a benchmarking framework to rigorously evaluate both per-modality image synthesis and the multi-modal coherence of reconstructed scenes.