GeoNet: Unsupervised Learning of Dense Depth, Optical Flow and Camera Pose

TL;DR

GeoNet is an unsupervised framework that jointly learns dense depth, optical flow, and camera pose from videos by leveraging 3D scene geometry, achieving state-of-the-art results without supervision.

Contribution

It introduces a unified end-to-end unsupervised learning approach that couples depth, flow, and pose estimation using geometric relationships and an adaptive consistency loss.

Findings

Achieves state-of-the-art unsupervised results on KITTI dataset

Performs comparably with supervised methods in depth, flow, and pose estimation

Effectively handles occlusions and non-Lambertian regions

Abstract

We propose GeoNet, a jointly unsupervised learning framework for monocular depth, optical flow and ego-motion estimation from videos. The three components are coupled by the nature of 3D scene geometry, jointly learned by our framework in an end-to-end manner. Specifically, geometric relationships are extracted over the predictions of individual modules and then combined as an image reconstruction loss, reasoning about static and dynamic scene parts separately. Furthermore, we propose an adaptive geometric consistency loss to increase robustness towards outliers and non-Lambertian regions, which resolves occlusions and texture ambiguities effectively. Experimentation on the KITTI driving dataset reveals that our scheme achieves state-of-the-art results in all of the three tasks, performing better than previously unsupervised methods and comparably with supervised ones.