STN-Homography: estimate homography parameters directly

TL;DR

This paper presents STN-Homography, a CNN-based model that directly estimates homography matrices with high accuracy and efficiency, outperforming existing methods on the MSCOCO dataset.

Contribution

The paper introduces a novel hierarchical CNN architecture for direct homography estimation using coordinate normalization, improving accuracy and speed over prior approaches.

Findings

Significantly outperforms state-of-the-art on MSCOCO dataset

Achieves real-time processing with 4.87 ms for 1 stage

Hierarchical approach reduces estimation error effectively

Abstract

In this paper, we introduce the STN-Homography model to directly estimate the homography matrix between image pair. Different most CNN-based homography estimation methods which use an alternative 4-point homography parameterization, we use prove that, after coordinate normalization, the variance of elements of coordinate normalized $3\times3$ homography matrix is very small and suitable to be regressed well with CNN. Based on proposed STN-Homography, we use a hierarchical architecture which stacks several STN-Homography models and successively reduce the estimation error. Effectiveness of the proposed method is shown through experiments on MSCOCO dataset, in which it significantly outperforms the state-of-the-art. The average processing time of our hierarchical STN-Homography with 1 stage is only 4.87 ms on the GPU, and the processing time for hierarchical STN-Homography with 3 stages is 17.85 ms. The code will soon be open sourced.