MegaSaM: Accurate, Fast, and Robust Structure and Motion from Casual Dynamic Videos

TL;DR

MegaSaM introduces a deep visual SLAM system that accurately, quickly, and robustly estimates camera motion and depth from casual, dynamic monocular videos, outperforming prior methods in real-world scenarios.

Contribution

The paper adapts a deep visual SLAM framework for dynamic scenes with minimal parallax, achieving high accuracy and robustness in challenging real-world videos.

Findings

Outperforms prior methods in accuracy and robustness

Operates efficiently on complex dynamic videos

Handles unconstrained camera motions effectively

Abstract

We present a system that allows for accurate, fast, and robust estimation of camera parameters and depth maps from casual monocular videos of dynamic scenes. Most conventional structure from motion and monocular SLAM techniques assume input videos that feature predominantly static scenes with large amounts of parallax. Such methods tend to produce erroneous estimates in the absence of these conditions. Recent neural network-based approaches attempt to overcome these challenges; however, such methods are either computationally expensive or brittle when run on dynamic videos with uncontrolled camera motion or unknown field of view. We demonstrate the surprising effectiveness of a deep visual SLAM framework: with careful modifications to its training and inference schemes, this system can scale to real-world videos of complex dynamic scenes with unconstrained camera paths, including videos with little camera parallax. Extensive experiments on both synthetic and real videos demonstrate that our system is significantly more accurate and robust at camera pose and depth estimation when compared with prior and concurrent work, with faster or comparable running times. See interactive results on our project page: https://mega-sam.github.io/