Accelerating Transformer-Based Monocular SLAM via Geometric Utility Scoring

2026-04-09

AI Summary

This work addresses the excessive computational redundancy in monocular SLAM based on geometric foundation models (GFMs) within dense video streams, which arises from the need for dense geometric decoding. To mitigate this, the authors propose LeanGate—a lightweight feedforward frame gating network that predicts the geometric utility score of each frame prior to GFM feature extraction, enabling early filtering of redundant frames. LeanGate is the first method to perform utility assessment before geometric decoding, offering plug-and-play compatibility with existing GFM-SLAM pipelines. Experimental results demonstrate that LeanGate reduces tracking FLOPs by over 85% on standard SLAM benchmarks, achieves a 5× end-to-end throughput speedup, and maintains tracking and mapping accuracy comparable to dense baselines.

Technology Category

• Computer Vision: Motion & Tracking
• Machine Learning: Mixture of Experts (MoE)
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Security and Privacy: Large-scale security measurements

Abstract

Geometric Foundation Models (GFMs) have recently advanced monocular SLAM by providing robust, calibration-free 3D priors. However, deploying these models on dense video streams introduces significant computational redundancy. Current GFM-based SLAM systems typically rely on post hoc keyframe selection. Because of this, they must perform expensive dense geometric decoding simply to determine whether a frame contains novel geometry, resulting in late rejection and wasted computation. To mitigate this inefficiency, we propose LeanGate, a lightweight feed-forward frame-gating network. LeanGate predicts a geometric utility score to assess a frame's mapping value prior to the heavy GFM feature extraction and matching stages. As a predictive plug-and-play module, our approach bypasses over 90% of redundant frames. Evaluations on standard SLAM benchmarks demonstrate that LeanGate reduces tracking FLOPs by more than 85% and achieves a 5x end-to-end throughput speedup. Furthermore, it maintains the tracking and mapping accuracy of dense baselines.

Problem

• monocular SLAM
• Geometric Foundation Models
• computational redundancy
• keyframe selection
• dense video streams

Innovation

• Geometric Utility Scoring
• Frame Gating
• Monocular SLAM
• Geometric Foundation Models
• Computational Efficiency