Loss it right: Euclidean and riemannian metrics in learning-based visual odometry

Research output: Contribution to book/anthology/report/proceedingArticle in proceedingsResearchpeer-review

Abstract

This paper overviews different pose representations and metric functions in visual odometry (VO) networks. The performance of VO networks heavily relies on how their architecture encodes the information. The choice of pose representation and loss function significantly impacts network convergence and generalization. We investigate these factors in the VO network DeepVO by implementing loss functions based on Euler, quaternion, and chordal distance and analyzing their influence on performance. The results of this study provide insights into how loss functions affect the designing of efficient and accurate VO networks for camera motion estimation. The experiments illustrate that a distance that complies with the mathematical requirements of a metric, such as the chordal distance, provides better generalization and faster convergence. The code for the experiments can be found at https://github.com/remaro-network/ Loss_VO_right.

Original languageEnglish
Title of host publicationISR Europe 2023 : 56th International Symposium on Robotics, in cooperation with Fraunhofer IPA September 26 – 27, 2023 in Stuttgart
Number of pages5
PublisherVDE Verlag GmbH
Publication date2023
Pages107-111
ISBN (Print)978-3-8007-6140
ISBN (Electronic)978-3-8007-6141-8
Publication statusPublished - 2023
EventISR Europe 2023; 56th International Symposium on Robotics - Stuttgart, Germany
Duration: 27 Sept 2023 → …

Conference

ConferenceISR Europe 2023; 56th International Symposium on Robotics
Country/TerritoryGermany
CityStuttgart
Period27/09/2023 → …

Keywords

  • Deep Learning
  • Lie algebra
  • Riemannian geometry
  • Visual Odometry

Fingerprint

Dive into the research topics of 'Loss it right: Euclidean and riemannian metrics in learning-based visual odometry'. Together they form a unique fingerprint.

Cite this