relation: https://khub.utp.edu.my/scholars/14165/
title: WPO-net: Windowed pose optimization network for monocular visual odometry estimation
creator: Gadipudi, N.
creator: Elamvazuthi, I.
creator: Lu, C.-K.
creator: Paramasivam, S.
creator: Su, S.
description: Visual odometry is the process of estimating incremental localization of the camera in 3-dimensional space for autonomous driving. There have been new learning-based methods which do not require camera calibration and are robust to external noise. In this work, a new method that do not require camera calibration called the �windowed pose optimization network� is proposed to estimate the 6 degrees of freedom pose of a monocular camera. The architecture of the proposed network is based on supervised learning-based methods with feature encoder and pose regressor that takes multiple consecutive two grayscale image stacks at each step for training and enforces the composite pose constraints. The KITTI dataset is used to evaluate the performance of the proposed method. The proposed method yielded rotational error of 3.12 deg/100 m, and the training time is 41.32 ms, while inference time is 7.87 ms. Experiments demonstrate the competitive performance of the proposed method to other state-of-the-art related works which shows the novelty of the proposed technique. © 2021 by the authors. Licensee MDPI, Basel, Switzerland.
publisher: MDPI
date: 2021
type: Article
type: PeerReviewed
identifier:   Gadipudi, N. and Elamvazuthi, I. and Lu, C.-K. and Paramasivam, S. and Su, S.  (2021) WPO-net: Windowed pose optimization network for monocular visual odometry estimation.  Sensors, 21 (23).   ISSN 14248220     
relation: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85120608377&doi=10.3390%2fs21238155&partnerID=40&md5=5e16a9cc60c3443b97dce0fb426bc32e
relation: 10.3390/s21238155
identifier: 10.3390/s21238155