Event
Geometry-Aware Epipolar Correspondence Graph for Monocular Visual Odometry
1 reports · 1 sources
Prateeth Rao and Sachit Rao submitted a paper proposing a geometry-aware epipolar correspondence graph method for monocular visual odometry: first, an existing detector-free model is used to find matching corresponding points between two images; then, epipolar constraints are used to guide the construction of the topology of the correspondence graph; subsequently, a graph learning module generates geometry context vectors, which are passed through an MLP to output rotation and translation to estimate the camera pose. The authors state that the method was validated for monocular visual odometry applications on outdoor benchmark datasets such as KiTTi, ETH3D, King's College, and Tartan Air, and that the paper has been submitted to the journal Pattern Recognition.
Report timeline
- Geometry-Aware Epipolar Correspondence Graphs for Monocular Visual Odometry
Prateeth Rao and Sachit Rao propose a geometry-aware epipolar correspondence graph method that uses epipolar constraints to guide the construction and pruning of correspondence graphs, then a graph learning module generates a geometric context vector, which is passed through an MLP to output rotation and translation for camera pose estimation.
arXiv cs.RO ·