Content-Aware Unsupervised Deep Homography Estimation

TL;DR

This paper introduces an unsupervised deep learning approach for homography estimation that effectively handles depth disparities and moving objects by learning reliable regions and using a novel triplet loss, outperforming existing methods.

Contribution

The paper proposes a new architecture for unsupervised deep homography estimation that incorporates outlier masking and a specialized triplet loss, addressing real-world challenges.

Findings

Outperforms state-of-the-art methods in diverse scenes

Effectively handles depth disparities and moving objects

Uses learned deep features for loss calculation

Abstract

Homography estimation is a basic image alignment method in many applications. It is usually conducted by extracting and matching sparse feature points, which are error-prone in low-light and low-texture images. On the other hand, previous deep homography approaches use either synthetic images for supervised learning or aerial images for unsupervised learning, both ignoring the importance of handling depth disparities and moving objects in real world applications. To overcome these problems, in this work we propose an unsupervised deep homography method with a new architecture design. In the spirit of the RANSAC procedure in traditional methods, we specifically learn an outlier mask to only select reliable regions for homography estimation. We calculate loss with respect to our learned deep features instead of directly comparing image content as did previously. To achieve the unsupervised training, we also formulate a novel triplet loss customized for our network. We verify our method by conducting comprehensive comparisons on a new dataset that covers a wide range of scenes with varying degrees of difficulties for the task. Experimental results reveal that our method outperforms the state-of-the-art including deep solutions and feature-based solutions.