Self-Supervised Deep Pose Corrections for Robust Visual Odometry

@article{Wagstaff2020SelfSupervisedDP,
  title={Self-Supervised Deep Pose Corrections for Robust Visual Odometry},
  author={Brandon Wagstaff and Valentin Peretroukhin and Jonathan Kelly},
  journal={2020 IEEE International Conference on Robotics and Automation (ICRA)},
  year={2020},
  pages={2331-2337}
}
We present a self-supervised deep pose correction (DPC) network that applies pose corrections to a visual odom-etry estimator to improve its accuracy. Instead of regressing inter-frame pose changes directly, we build on prior work that uses data-driven learning to regress pose corrections that account for systematic errors due to violations of modelling assumptions. Our self-supervised formulation removes any requirement for six-degrees-of-freedom ground truth and, in contrast to expectations… Expand
4 Citations

Figures and Tables from this paper

Self-Supervised Learning of Visual Odometry
  • Lesheng Song, W. Luo
  • Computer Science
  • 2020 International Conference on Information Science, Parallel and Distributed Systems (ISPDS)
  • 2020
Robot Perception enables Complex Navigation Behavior via Self-Supervised Learning
  • PDF
Deep Online Correction for Monocular Visual Odometry
  • Jiaxin Zhang, Wei Sui, Xinggang Wang, Wenming Meng, Hongmei Zhu, Qian Zhang
  • Computer Science
  • ArXiv
  • 2021
  • PDF
Self-Supervised Scale Recovery for Monocular Depth and Egomotion Estimation
  • 2
  • PDF

References

SHOWING 1-10 OF 35 REFERENCES
Geometric Consistency for Self-Supervised End-to-End Visual Odometry
  • 32
  • PDF
DPC-Net: Deep Pose Correction for Visual Localization
  • 31
  • PDF
SfMLearner++: Learning Monocular Depth & Ego-Motion Using Meaningful Geometric Constraints
  • 13
  • PDF
Unsupervised Learning of Monocular Depth Estimation and Visual Odometry with Deep Feature Reconstruction
  • 285
  • PDF
UnDeepVO: Monocular Visual Odometry Through Unsupervised Deep Learning
  • 264
  • PDF
Unsupervised Learning of Depth and Ego-Motion from Video
  • 1,156
  • Highly Influential
  • PDF
DeepVO: Towards end-to-end visual odometry with deep Recurrent Convolutional Neural Networks
  • 324
  • PDF
Direct Sparse Odometry
  • 1,154
  • PDF
...
1
2
3
4
...