Transfer Learning from Synthetic Data in the Camera Pose Estimation Problem

Jorge L. Charco, Jorge L. Charco, Angel D. Sappa, Angel D. Sappa, Boris X. Vintimilla, Henry O. Velesaca

2020

Abstract

This paper presents a novel Siamese network architecture, as a variant of Resnet-50, to estimate the relative camera pose on multi-view environments. In order to improve the performance of the proposed model a transfer learning strategy, based on synthetic images obtained from a virtual-world, is considered. The transfer learning consists of first training the network using pairs of images from the virtual-world scenario considering different conditions (i.e., weather, illumination, objects, buildings, etc.); then, the learned weight of the network are transferred to the real case, where images from real-world scenarios are considered. Experimental results and comparisons with the state of the art show both, improvements on the relative pose estimation accuracy using the proposed model, as well as further improvements when the transfer learning strategy (synthetic-world data transfer learning real-world data) is considered to tackle the limitation on the training due to the reduced number of pairs of real-images on most of the public data sets.

Paper Citation

in Harvard Style

Charco J., Sappa A., Vintimilla B. and Velesaca H. (2020). Transfer Learning from Synthetic Data in the Camera Pose Estimation Problem. In Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2020) - Volume 4: VISAPP; ISBN 978-989-758-402-2, SciTePress, pages 498-505. DOI: 10.5220/0009167604980505

in Bibtex Style

@conference{visapp20,
author={Jorge L. Charco and Angel D. Sappa and Boris X. Vintimilla and Henry O. Velesaca},
title={Transfer Learning from Synthetic Data in the Camera Pose Estimation Problem},
booktitle={Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2020) - Volume 4: VISAPP},
year={2020},
pages={498-505},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0009167604980505},
isbn={978-989-758-402-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2020) - Volume 4: VISAPP
TI - Transfer Learning from Synthetic Data in the Camera Pose Estimation Problem
SN - 978-989-758-402-2
AU - Charco J.
AU - Sappa A.
AU - Vintimilla B.
AU - Velesaca H.
PY - 2020
SP - 498
EP - 505
DO - 10.5220/0009167604980505
PB - SciTePress