Deep Fundamental Matrix Estimation without Correspondences

TL;DR

This paper introduces neural network architectures that estimate fundamental matrices directly from image pairs without relying on key-point correspondences, enabling robust performance even with occlusions or large viewpoint changes.

Contribution

The paper presents novel neural network models that preserve the mathematical properties of fundamental matrices and operate end-to-end without correspondence estimation.

Findings

Achieves competitive performance on KITTI dataset

Handles large occlusions and viewpoint variations effectively

Eliminates the need for key-point correspondence extraction

Abstract

Estimating fundamental matrices is a classic problem in computer vision. Traditional methods rely heavily on the correctness of estimated key-point correspondences, which can be noisy and unreliable. As a result, it is difficult for these methods to handle image pairs with large occlusion or significantly different camera poses. In this paper, we propose novel neural network architectures to estimate fundamental matrices in an end-to-end manner without relying on point correspondences. New modules and layers are introduced in order to preserve mathematical properties of the fundamental matrix as a homogeneous rank-2 matrix with seven degrees of freedom. We analyze performance of the proposed models using various metrics on the KITTI dataset, and show that they achieve competitive performance with traditional methods without the need for extracting correspondences.