Loss it right: Euclidean and Riemannian Metrics in Learning-based Visual Odometry

Conference: ISR Europe 2023 - 56th International Symposium on Robotics
09/26/2023 - 09/27/2023 at Stuttgart, Germany

Proceedings: ISR Europe 2023

Pages: 5Language: englishTyp: PDF

Authors:
Alvarez-Tunon, Olaya (Artificial Intelligence in Robotics Laboratory (AiRLab), the Department of Electrical and Computer Engineering, Aarhus University, Denmark)
Brodskiy, Yury (EIVA a/s, Skanderborg, Denmark)
Kayacan, Erdal (Automatic Control Group (RAT), Paderborn University, Paderborn, Germany)

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.