SRSplat: Feed-Forward Super-Resolution Gaussian Splatting from Sparse Multi-View Images

TL;DR

SRSplat is a feed-forward framework that reconstructs high-resolution 3D scenes from sparse low-resolution images by leveraging external reference images and internal texture cues, significantly improving detail recovery.

Contribution

The paper introduces SRSplat, a novel method combining external references and internal cues with modules RGFE and TADC for enhanced 3D reconstruction from LR images.

Findings

Outperforms existing methods on multiple datasets.

Demonstrates strong cross-dataset generalization.

Effectively recovers fine texture details.

Abstract

Feed-forward 3D reconstruction from sparse, low-resolution (LR) images is a crucial capability for real-world applications, such as autonomous driving and embodied AI. However, existing methods often fail to recover fine texture details. This limitation stems from the inherent lack of high-frequency information in LR inputs. To address this, we propose \textbf{SRSplat}, a feed-forward framework that reconstructs high-resolution 3D scenes from only a few LR views. Our main insight is to compensate for the deficiency of texture information by jointly leveraging external high-quality reference images and internal texture cues. We first construct a scene-specific reference gallery, generated for each scene using Multimodal Large Language Models (MLLMs) and diffusion models. To integrate this external information, we introduce the \textit{Reference-Guided Feature Enhancement (RGFE)} module, which aligns and fuses features from the LR input images and their reference twin image. Subsequently, we train a decoder to predict the Gaussian primitives using the multi-view fused feature obtained from \textit{RGFE}. To further refine predicted Gaussian primitives, we introduce \textit{Texture-Aware Density Control (TADC)}, which adaptively adjusts Gaussian density based on the internal texture richness of the LR inputs. Extensive experiments demonstrate that our SRSplat outperforms existing methods on various datasets, including RealEstate10K, ACID, and DTU, and exhibits strong cross-dataset and cross-resolution generalization capabilities.