Flash-Mono: Feed-Forward Accelerated Gaussian Splatting Monocular SLAM

2026-04-03

AI Summary

This work addresses the inefficiency, insufficient geometric accuracy, and multi-view inconsistency inherent in monocular Gaussian splatting SLAM due to per-frame optimization. To overcome these limitations, we propose the first feedforward monocular Gaussian splatting SLAM system, which employs a recurrent feedforward frontend to aggregate multi-frame features and directly predict camera poses and Gaussian attributes. We introduce 2D Gaussian surfels to enhance geometric fidelity and utilize a latent state as a submap descriptor to enable efficient loop closure detection and Sim(3) global optimization. The proposed method achieves a tenfold speedup over conventional optimization-based approaches while attaining state-of-the-art performance in tracking accuracy and reconstruction quality, making it well-suited for real-time 3D reconstruction and embodied intelligence applications.

Abstract

Monocular 3D Gaussian Splatting SLAM suffers from critical limitations in time efficiency, geometric accuracy, and multi-view consistency. These issues stem from the time-consuming $\textit{Train-from-Scratch}$ optimization and the lack of inter-frame scale consistency from single-frame geometry priors. We contend that a feed-forward paradigm, leveraging multi-frame context to predict Gaussian attributes directly, is crucial for addressing these challenges. We present Flash-Mono, a system composed of three core modules: a feed-forward prediction frontend, a 2D Gaussian Splatting mapping backend, and an efficient hidden-state-based loop closure module. We trained a recurrent feed-forward frontend model that progressively aggregates multi-frame visual features into a hidden state via cross attention and jointly predicts camera poses and per-pixel Gaussian properties. By directly predicting Gaussian attributes, our method bypasses the burdensome per-frame optimization required in optimization-based GS-SLAM, achieving a $\textbf{10x}$ speedup while ensuring high-quality rendering. The power of our recurrent architecture extends beyond efficient prediction. The hidden states act as compact submap descriptors, facilitating efficient loop closure and global $\mathrm{Sim}(3)$ optimization to mitigate the long-standing challenge of drift. For enhanced geometric fidelity, we replace conventional 3D Gaussian ellipsoids with 2D Gaussian surfels. Extensive experiments demonstrate that Flash-Mono achieves state-of-the-art performance in both tracking and mapping quality, highlighting its potential for embodied perception and real-time reconstruction applications. Project page: https://victkk.github.io/flash-mono.

Problem

• Monocular SLAM
• Gaussian Splatting
• Time Efficiency
• Geometric Accuracy
• Multi-view Consistency

Innovation

• Feed-forward SLAM
• Gaussian Splatting
• Monocular Reconstruction
• Loop Closure
• 2D Gaussian Surfels