Is Geometry Enough for Matching in Visual Localization?

TL;DR

This paper introduces GoMatch, a geometry-only approach for visual localization that reduces storage needs and privacy concerns while maintaining high accuracy, by using a novel bearing vectors representation of 3D points.

Contribution

GoMatch presents a new geometry-based matching method that eliminates the need for visual descriptors, addressing storage and privacy issues in large-scale localization.

Findings

Achieves comparable accuracy to visual descriptor methods with significantly less storage.

Reduces median pose errors by over 10 meters and 95 degrees on benchmark datasets.

Demonstrates feasibility for city-scale localization without visual descriptor storage.

Abstract

In this paper, we propose to go beyond the well-established approach to vision-based localization that relies on visual descriptor matching between a query image and a 3D point cloud. While matching keypoints via visual descriptors makes localization highly accurate, it has significant storage demands, raises privacy concerns and requires update to the descriptors in the long-term. To elegantly address those practical challenges for large-scale localization, we present GoMatch, an alternative to visual-based matching that solely relies on geometric information for matching image keypoints to maps, represented as sets of bearing vectors. Our novel bearing vectors representation of 3D points, significantly relieves the cross-modal challenge in geometric-based matching that prevented prior work to tackle localization in a realistic environment. With additional careful architecture design, GoMatch improves over prior geometric-based matching work with a reduction of (10.67m,95.7deg) and (1.43m, 34.7deg) in average median pose errors on Cambridge Landmarks and 7-Scenes, while requiring as little as 1.5/1.7% of storage capacity in comparison to the best visual-based matching methods. This confirms its potential and feasibility for real-world localization and opens the door to future efforts in advancing city-scale visual localization methods that do not require storing visual descriptors.