Appearance Decomposition Gaussian Splatting for Multi-Traversal Reconstruction

2026-04-07

AI Summary

This work addresses the challenge of appearance inconsistency in multi-pass scene reconstruction caused by varying illumination and environmental conditions. The authors propose the ADM-GS framework, which achieves appearance disentanglement while preserving a shared geometric structure. ADM-GS introduces, for the first time in Gaussian splatting, an explicit appearance decomposition mechanism that models static backgrounds using illumination-invariant material properties and illumination-dependent components. By integrating neural light fields, surface normals, and reflection vectors, the method effectively separates diffuse and specular reflectance. A frequency-separated hybrid encoding strategy combined with geometry-guided Gaussian splatting significantly enhances cross-pass appearance consistency, yielding a 0.98 dB improvement in PSNR on the Argoverse 2 and Waymo Open datasets.

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.

Problem

• multi-traversal reconstruction
• appearance inconsistency
• illumination variation
• static background
• scene reconstruction

Innovation

• Appearance Decomposition
• Gaussian Splatting
• Neural Light Field
• Multi-Traversal Reconstruction
• Frequency-Separated Encoding