MegaSaM: Accurate, Fast, and Robust Structure and Motion from Casual Dynamic Videos
This addresses the challenge of structure and motion estimation in dynamic videos for applications like robotics and augmented reality, representing a strong specific gain rather than a foundational breakthrough.
The paper tackles the problem of estimating camera parameters and depth maps from casual monocular videos of dynamic scenes, where conventional methods fail, and demonstrates that their system achieves significantly more accurate and robust results compared to prior work, with faster or comparable running times.
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/