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

TY - GEN

T1 - Loss it right

T2 - ISR Europe 2023; 56th International Symposium on Robotics

AU - Alvarez Tunon, Olaya

AU - Brodskiy, Yury

AU - Kayacan, Erdal

PY - 2023

Y1 - 2023

N2 - 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.

AB - 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.

KW - Deep Learning

KW - Lie algebra

KW - Riemannian geometry

KW - Visual Odometry

UR - http://www.scopus.com/inward/record.url?scp=85184348481&partnerID=8YFLogxK

M3 - Article in proceedings

SN - 978-3-8007-6140

SP - 107

EP - 111

BT - ISR Europe 2023

PB - VDE Verlag GmbH

Y2 - 27 September 2023

ER -