Photometric Visual Gyroscope for Full-View Spherical Camera

Antoine N. André, Guillaume Caron; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2022, pp. 5232-5235

Abstract

This article presents a visual gyroscope based on the spherical representation of dual-fisheye cameras. By using the fully available view of the scene thanks to a dual-fisheye camera and projecting the photometric information on a sphere, a highly precise orientation estimation along with a great convergence domain can be achieved. This method is validated with the help of images taken from the PanoraMIS dataset to evaluate quantitatively the performances.

Related Material

[pdf]
[bibtex]
@InProceedings{Andre_2022_CVPR, author = {Andr\'e, Antoine N. and Caron, Guillaume}, title = {Photometric Visual Gyroscope for Full-View Spherical Camera}, booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops}, month = {June}, year = {2022}, pages = {5232-5235} }