Self-Supervised 3D Keypoint Learning for Ego-motion Estimation
This addresses the need for accurate ego-motion estimation in autonomous vehicles by improving generalization to real-world 3D scenes, though it builds incrementally on existing self-supervised and geometric methods.
The paper tackles the problem of learning robust 3D keypoints for ego-motion estimation in non-planar scenes with illumination variations, proposing a self-supervised method that jointly learns keypoint and depth estimation from unlabeled videos, achieving integration into visual odometry frameworks for autonomous vehicles.
Detecting and matching robust viewpoint-invariant keypoints is critical for visual SLAM and Structure-from-Motion. State-of-the-art learning-based methods generate training samples via homography adaptation to create 2D synthetic views with known keypoint matches from a single image. This approach, however, does not generalize to non-planar 3D scenes with illumination variations commonly seen in real-world videos. In this work, we propose self-supervised learning of depth-aware keypoints directly from unlabeled videos. We jointly learn keypoint and depth estimation networks by combining appearance and geometric matching via a differentiable structure-from-motion module based on Procrustean residual pose correction. We describe how our self-supervised keypoints can be integrated into state-of-the-art visual odometry frameworks for robust and accurate ego-motion estimation of autonomous vehicles in real-world conditions.