Research Analyzer
← Back ICRA 2026

EC3R-SLAM: Efficient and Consistent Monocular Dense SLAM with Feed-Forward 3D Reconstruction

lingxiang hu, Naima Ait Oufroukh, Fabien Bonardi, Ghandour Raymond

PDF

AI summary

Key figure (auto-extracted from paper)
EC3R-SLAM delivers real-time, calibration-free dense mapping on consumer hardware by coupling lightweight tracking with feed-forward 3D reconstruction and loop closure.
Monocular SLAM Dense Reconstruction Feed-Forward Model Calibration-Free Real-Time Mapping Loop Closure

Problem

Existing monocular dense SLAM methods suffer from high GPU memory consumption, significant latency, and reliance on camera calibration, which hinder real-time deployment on resource-constrained platforms.

Approach

The system tightly couples a lightweight feature-based tracking module with a feed-forward 3D reconstruction model, using local and global loop closures to enforce multi-view consistency while jointly estimating camera intrinsics.

Key results

  • Operates at >30 FPS with under 10 GB GPU memory
  • Achieves competitive accuracy across TUM-RGBD, Replica, and 7-Scenes benchmarks
  • Enables calibration-free operation by jointly estimating camera intrinsics
  • Reduces accumulated drift via Sim(3) pose graph optimization and point correction

Why it matters

Enables real-time, resource-efficient dense mapping and localization for robotics and AR/VR applications on consumer-grade hardware without requiring camera calibration.

Abstract

The application of monocular dense Simultane- ous Localization and Mapping (SLAM) is often hindered by high latency, large GPU memory consumption, and reliance on camera calibration. To relax this constraint, we propose EC3R-SLAM, a novel calibration-free monocular dense SLAM framework that jointly achieves high localization and mapping accuracy, low latency, and low GPU memory consumption. This enables the framework to achieve efficiency through the coupling of a tracking module, which maintains a sparse map of feature points, and a mapping module based on a feed- forward 3D reconstruction model that simultaneously estimates camera intrinsics. In addition, both local and global loop closures are incorporated to ensure mid-term and long-term data association, enforcing multi-view consistency and thereby enhancing the overall accuracy and robustness of the system. Experiments across multiple benchmarks show that EC3R- SLAM achieves competitive performance compared to state-of- the-art methods, while being faster and more memory-efficient. Moreover, it runs effectively even on resource-constrained platforms such as laptops and Jetson Orin NX, highlighting its potential for real-world robotics applications. Project page: https://h0xg.github.io/ec3r/

Index terms

SLAM Mapping Localization

Related papers