Appearance Decomposition Gaussian Splatting for Multi-Traversal Reconstruction

Yangyi Xiao; Siting Zhu; Baoquan Yang; Tianchen Deng; Yongbo Chen; Hesheng Wang

cs.CVApr 7, 2026

Appearance Decomposition Gaussian Splatting for Multi-Traversal Reconstruction

Authors: Yangyi Xiao, Siting Zhu, Baoquan Yang, Tianchen Deng, Yongbo Chen, Hesheng Wang

Organizations: Department of Automation, Key Laboratory of System Control and Information Processing of Ministry of Education, State Key Laboratory of Avionics Integration and Aviation System-of-Systems Synthesis, Shanghai Jiao Tong University, Shanghai 200240, China.

Abstract

Multi-traversal scene reconstruction is important for high-fidelity autonomous driving simulation and digital twin construction. This task involves integrating multiple sequences captured from the same geographical area at different times. In this context, a primary challenge is the significant appearance inconsistency across traversals caused by varying illumination and environmental conditions, despite the shared underlying geometry. This paper presents ADM-GS (Appearance Decomposition Gaussian Splatting for Multi-Traversal Reconstruction), a framework that applies an explicit appearance decomposition to the static background to alleviate appearance entanglement across traversals. For the static background, we decompose the appearance into traversal-invariant material, representing intrinsic material properties, and traversal-dependent illumination, capturing lighting variations. Specifically, we propose a neural light field that utilizes a frequency-separated hybrid encoding strategy. By incorporating surface normals and explicit reflection vectors, this design separately captures low-frequency diffuse illumination and high-frequency specular reflections. Quantitative evaluations on the Argoverse 2 and Waymo Open datasets demonstrate the effectiveness of ADM-GS. In multi-traversal experiments, our method achieves a +0.98 dB PSNR improvement over existing latent-based baselines while producing more consistent appearance across traversals. Code will be available at https://github.com/IRMVLab/ADM-GS.

Figures & tables

Fig. 1: Consistent Multi-Traversal Reconstruction via Appearance Decomposition. (Left) Input images from multi-traversal scenes exhibit appearance inconsistencies due to changes in illumination, weather, and time of day. (Middle) ADM-GS decomposes the static scene appearance into a traversal-invariant material field ( $M$ ) and a traversal-dependent light field ( $L_{m}$ ), enabling structured modeling of cross-traversal appearance variation. (Right) Rendering results show that ADM-GS preserves consistent material across different environmental conditions while capturing traversal-specific illumination variations, leading to improved multi-traversal novel view synthesis.

Fig. 2: Framework of Appearance Decomposition Gaussian Splatting for Multi-Traversal Reconstruction. ADM-GS represents the scene with a hybrid scene graph consisting of static, object, and sky nodes. Its core static node decomposes appearance into a traversal-invariant material field and a traversal-dependent light field, where normal and reflection-vector cues are introduced to improve illumination prediction. A traversal-conditioned opacity gating module, adapted from 4DGF [ 1 ] , uses the traversal embedding $e_{m}$ to predict traversal-specific opacity, which suppresses geometrically inconsistent regions during training. Ground-truth RGB images and pseudo supervision on material and normals are used for optimization.

Fig. 3: Illustration of frequency-separated illumination cues. Surface normals are mainly associated with low-frequency diffuse illumination, while reflection vectors are more informative for high-frequency view-dependent specular effects.

Dataset	Method	Novel View Synthesis			Scene Reconstruction
Dataset	Method	PSNR $\uparrow$	SSIM $\uparrow$	LPIPS $\downarrow$	PSNR $\uparrow$	SSIM $\uparrow$	LPIPS $\downarrow$
Argoverse 2	PVG [ 28 ]	23.04	0.782	0.290	24.37	0.824	0.275
	StreetGS [ 15 ]	23.40	0.795	0.197	25.02	0.857	0.181
	OmniRe [ 18 ]	23.61	0.797	0.194	25.31	0.859	0.176
	Bilateral-Driving [ 21 ]	23.70	0.794	0.183	25.74	0.865	0.164
	4DGF [ 1 ]	29.48	0.890	0.155	30.17	0.902	0.146

TABLE I: Quantitative comparison on single-traversal scenarios. We report results for novel view synthesis and scene reconstruction. Bold and underlined values indicate the best and second-best results, respectively.

Panel of Figure 5 — Fig. 4: Visual comparison of novel view synthesis on dynamic scenes. We compare our method with recent approaches on the Argoverse 2 (top row) and Waymo Open (bottom row) datasets. Methods such as Bilateral-Driving [ 21 ] and OmniRe [ 18 ] show noticeable blur and ghosting artifacts in dynamic-object regions. In contrast, our method produces sharper novel-view renderings and is visually closer to the ground truth.

Method	PSNR $\uparrow$	SSIM $\uparrow$	LPIPS $\downarrow$
4DGF [ 1 ]	26.40	0.780	0.388
MTGS [ 2 ]	23.28	0.717	0.287
ADM-GS (Ours)	27.38	0.788	0.368

TABLE II: Quantitative comparison on multi-traversal reconstruction on the Argoverse 2 dataset. The models are trained on three traversals with varying illumination conditions. Bold indicates the best performance.

Panel of Figure 7 — Fig. 5: Qualitative analysis of appearance decomposition on the Argoverse 2 multi-traversal dataset. Each row corresponds to a different traversal of the same scene. The estimated material remains largely consistent across traversals, while the illumination maps capture traversal-dependent illumination changes. The normal maps provide geometric cues that support illumination estimation.

Panel of Figure 8 — Fig. 6: Qualitative application: cross-traversal scene relighting. We demonstrate the controllability of the learned decomposition by transferring traversal-dependent illumination from Traversal $T_{2}$ to a scene observed in Traversal $T_{1}$ . Column 1 shows the source image from $T_{1}$ . Column 2 shows the target traversal used to provide the illumination condition from $T_{2}$ . Column 3 shows the estimated traversal-invariant material map $\hat{\mathbf{M}}_{T_{1}}$ . Column 4 shows the relit result obtained by combining the material from $T_{1}$ with the illumination condition from $T_{2}$ . Observe that the shadow cast by the tree (red box) follows the illumination pattern of $T_{2}$ , while the underlying road texture remains consistent with $T_{1}$ . The relighting operation is applied only to the static background, while dynamic foreground objects are not edited.

Method Configuration	PSNR $\uparrow$	SSIM $\uparrow$	LPIPS $\downarrow$
Implicit Appearance	27.10	0.782	0.383
Explicit Decomposition	27.31	0.785	0.380
+ Geometry Constraints	27.39	0.789	0.371

TABLE III: System-level ablation on Argoverse 2. We analyze the transition from implicit appearance modeling to explicit appearance decomposition on the multi-traversal NVS task. Explicit Decomposition includes pseudo material supervision, and + Geometry Constraints further adds pseudo normal supervision together with geometric cues.

Variant	$\mathcal{L}_{normal}$	Inputs		Metrics
Variant	$\mathcal{L}_{normal}$	$r$	$n$	PSNR $\uparrow$	SSIM $\uparrow$	LPIPS $\downarrow$
Baseline	$\times$	$\times$	$\times$	28.65	0.856	0.211
+ Normal Loss	✓	$\times$	$\times$	28.77	0.859	0.196
+ Reflection	✓	✓	$\times$	28.86	0.861	0.187
+ Refl. & Normal	✓	✓	✓	28.91	0.862	0.186

TABLE IV: Ablation of input features for the illumination MLP on Waymo. We analyze the impact of geometric supervision (

\mathcal{L}_{normal}

) and different input feature combinations (

r,n

) on the unseen test set.

Explore similar work

SpectralSplat: Appearance-Disentangled Feed-Forward Gaussian Splatting for Driving Scenes
Apr 3, 2026Quentin Herau, Tianshuo Xu, Depu Meng +53D Gaussian Splatting Representation Disentanglement
Applied Intuition · The Hong Kong University of Science and Technology · University of California, Berkeley
ReconDrive: Fast Feed-Forward 4D Gaussian Splatting for Autonomous Driving Scene Reconstruction
Mar 8, 2026Haibao Yu, Kuntao Xiao, Jiahang Wang +7Dynamic Scene Reconstruction 3D Gaussian Splatting
The University of Hong Kong · Tuojing Intelligence · King’s College London +2
L2D2-GS: Learning to Densify for Feedforward Dynamic Gaussian Scene Reconstruction
Jun 28, 2026Zetian Song, Chenming Wu, Junnan Liu +63D Gaussian Splatting Adaptive Gaussian Densification
State Key Laboratory of Multimedia Information Processing, School of Computer Science, Peking University, Beijing, 100871, China · Xiaomi EV, Beijing, 100085, China

Apr 3, 2026J/K move · Enter open · S save

SpectralSplat: Appearance-Disentangled Feed-Forward Gaussian Splatting for Driving Scenes

Quentin Herau, Tianshuo Xu, Depu Meng, Jiezhi Yang+4

Applied Intuition · The Hong Kong University of Science and Technology · University of California, Berkeley

3D Gaussian Splatting Representation Disentanglement Feed-Forward 3D Reconstruction Autonomous Driving Perception

Open paper AskPDF arXiv

Feed-forward 3D Gaussian Splatting methods have achieved impressive reconstruction quality for autonomous driving scenes, yet they entangle scene geometry with transient appearance properties such as lighting, weather, and time of day. This coupling prevents relighting, appearance transfer, and consistent rendering across multi-traversal data captured under varying environmental conditions. We present SpectralSplat, a method that disentangles appearance from geometry within a feed-forward Gaussian Splatting framework. Our key insight is to factor color prediction into an appearance-agnostic base stream and and appearance-conditioned adapted stream, both produced by a shared MLP conditioned on a global appearance embedding derived from DINOv2 features. To enforce disentanglement, we train with paired observations generated by a hybrid relighting pipeline that combines physics-based intrinsic decomposition with diffusion based generative refinement, and supervise with complementary consistency, reconstruction, cross-appearance, and base color losses. We further introduce an appearance-adaptable temporal history that stores appearance-agnostic features, enabling accumulated Gaussians to be re-rendered under arbitrary target appearances. Experiments demonstrate that SpectralSplat preserves the reconstruction quality of the underlying backbone while enabling controllable appearance transfer and temporally consistent relighting across driving sequences.

Abstract

Explore similar work

SpectralSplat: Appearance-Disentangled Feed-Forward Gaussian Splatting for Driving Scenes

ReconDrive: Fast Feed-Forward 4D Gaussian Splatting for Autonomous Driving Scene Reconstruction

L2D2-GS: Learning to Densify for Feedforward Dynamic Gaussian Scene Reconstruction