Perceptual Loss for Robust Unsupervised Homography Estimation

TL;DR

This paper introduces biHomE, a novel unsupervised homography estimation method that uses a perceptual loss in feature space, demonstrating robustness to illumination changes and achieving state-of-the-art results.

Contribution

The paper proposes biHomE, a new unsupervised homography estimation approach that decouples representation learning from homography prediction using a fixed feature extractor.

Findings

biHomE achieves state-of-the-art performance on COCO dataset.

The method is robust to illumination variations.

It outperforms existing approaches in accuracy and robustness.

Abstract

Homography estimation is often an indispensable step in many computer vision tasks. The existing approaches, however, are not robust to illumination and/or larger viewpoint changes. In this paper, we propose bidirectional implicit Homography Estimation (biHomE) loss for unsupervised homography estimation. biHomE minimizes the distance in the feature space between the warped image from the source viewpoint and the corresponding image from the target viewpoint. Since we use a fixed pre-trained feature extractor and the only learnable component of our framework is the homography network, we effectively decouple the homography estimation from representation learning. We use an additional photometric distortion step in the synthetic COCO dataset generation to better represent the illumination variation of the real-world scenarios. We show that biHomE achieves state-of-the-art performance on synthetic COCO dataset, which is also comparable or better compared to supervised approaches. Furthermore, the empirical results demonstrate the robustness of our approach to illumination variation compared to existing methods.