WGANVO: Monocular Visual Odometry based on Generative Adversarial Networks
Published in
arXiv.org(2020)External Links:
TL
TL;DR
This work presents WGANVO, a Deep Learning based monocular Visual Odometry method, where a neural network is trained to regress a pose estimate from an image pair using a semi-supervised approach.
Abstract
In this work we present WGANVO, a Deep Learning based monocular Visual Odometry method. In particular, a neural network is trained to regress a pose estimate from an image pair. The training is performed using a semi-supervised approach. Unlike geometry based monocular methods, the proposed method can recover the absolute scale of the scene without neither prior knowledge nor extra information. The evaluation of the system is carried out on the well-known KITTI dataset where it is shown to work in real time and the accuracy obtained is encouraging to continue the development of Deep Learning based methods.
References34 items
1
Generative Adversarial Networks
2
CAM-Convs: Camera-Aware Multi-Scale Convolutions for Single-View Depth
3
GANVO: Unsupervised Deep Monocular Visual Odometry and Depth Estimation with Generative Adversarial Networks
4
Deep Virtual Stereo Odometry: Leveraging Deep Depth Prediction for Monocular Direct Sparse Odometry
5