WGANVO: Monocular Visual Odometry based on Generative Adversarial Networks
This addresses the scale ambiguity problem in monocular visual odometry for robotics and autonomous systems, but it is incremental as it builds on existing deep learning approaches.
The authors tackled monocular visual odometry by training a neural network to regress pose from image pairs, enabling absolute scale recovery without prior knowledge. The method achieved real-time performance on the KITTI dataset with encouraging accuracy.
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.