SCPNet: Unsupervised Cross-modal Homography Estimation via Intra-modal Self-supervised Learning

TL;DR

SCPNet introduces an unsupervised framework for cross-modal homography estimation using intra-modal self-supervised learning, correlation, and feature projection, achieving state-of-the-art results on satellite-map datasets without supervision.

Contribution

The paper presents the first effective unsupervised method for cross-modal homography estimation, leveraging intra-modal self-supervised learning and correlation-based architecture.

Findings

Outperforms supervised methods on satellite-map datasets by 14% MACE.

Achieves state-of-the-art unsupervised performance on multiple datasets.

Reduces mean average corner error significantly compared to prior approaches.

Abstract

We propose a novel unsupervised cross-modal homography estimation framework based on intra-modal Self-supervised learning, Correlation, and consistent feature map Projection, namely SCPNet. The concept of intra-modal self-supervised learning is first presented to facilitate the unsupervised cross-modal homography estimation. The correlation-based homography estimation network and the consistent feature map projection are combined to form the learnable architecture of SCPNet, boosting the unsupervised learning framework. SCPNet is the first to achieve effective unsupervised homography estimation on the satellite-map image pair cross-modal dataset, GoogleMap, under [-32,+32] offset on a 128x128 image, leading the supervised approach MHN by 14.0% of mean average corner error (MACE). We further conduct extensive experiments on several cross-modal/spectral and manually-made inconsistent datasets, on which SCPNet achieves the state-of-the-art (SOTA) performance among unsupervised approaches, and owns 49.0%, 25.2%, 36.4%, and 10.7% lower MACEs than the supervised approach MHN. Source code is available at https://github.com/RM-Zhang/SCPNet.