Camera Relocalization by Computing Pairwise Relative Poses Using Convolutional Neural Network

TL;DR

This paper introduces a scalable CNN-based camera relocalization method that estimates relative poses between images, enabling scene-independent localization without scene-specific training, and demonstrates its effectiveness on new and standard datasets.

Contribution

The paper presents a scene-agnostic CNN approach for camera relocalization that estimates relative poses and fuses them for accurate localization, without requiring scene-specific training.

Findings

Method generalizes well to unseen scenes

Outperforms recent CNN-based localization methods

Effective on both new dataset and 7 Scenes benchmark

Abstract

We propose a new deep learning based approach for camera relocalization. Our approach localizes a given query image by using a convolutional neural network (CNN) for first retrieving similar database images and then predicting the relative pose between the query and the database images, whose poses are known. The camera location for the query image is obtained via triangulation from two relative translation estimates using a RANSAC based approach. Each relative pose estimate provides a hypothesis for the camera orientation and they are fused in a second RANSAC scheme. The neural network is trained for relative pose estimation in an end-to-end manner using training image pairs. In contrast to previous work, our approach does not require scene-specific training of the network, which improves scalability, and it can also be applied to scenes which are not available during the training of the network. As another main contribution, we release a challenging indoor localisation dataset covering 5 different scenes registered to a common coordinate frame. We evaluate our approach using both our own dataset and the standard 7 Scenes benchmark. The results show that the proposed approach generalizes well to previously unseen scenes and compares favourably to other recent CNN-based methods.