SAda-Net: A Self-Supervised Adaptive Stereo Estimation CNN For Remote Sensing Image Data

TL;DR

SAda-Net introduces a self-supervised CNN for stereo estimation in remote sensing images, effectively reducing the need for ground-truth data by iteratively refining disparity maps through self-adaptation and consistency checks.

Contribution

It presents a novel self-supervised adaptive stereo estimation CNN that improves disparity accuracy without requiring ground-truth data, specifically tailored for remote sensing applications.

Findings

Effective disparity map refinement through self-supervision

Reduced reliance on ground-truth data in remote sensing

Code publicly available for reproducibility

Abstract

Stereo estimation has made many advancements in recent years with the introduction of deep-learning. However the traditional supervised approach to deep-learning requires the creation of accurate and plentiful ground-truth data, which is expensive to create and not available in many situations. This is especially true for remote sensing applications, where there is an excess of available data without proper ground truth. To tackle this problem, we propose a self-supervised CNN with self-improving adaptive abilities. In the first iteration, the created disparity map is inaccurate and noisy. Leveraging the left-right consistency check, we get a sparse but more accurate disparity map which is used as an initial pseudo ground-truth. This pseudo ground-truth is then adapted and updated after every epoch in the training step of the network. We use the sum of inconsistent points in order to track the network convergence. The code for our method is publicly available at: https://github.com/thedodo/SAda-Net}{https://github.com/thedodo/SAda-Net