Benchmarking image retrieval for visual localization

Visual localization, i.e., camera pose estimation in a known scene, is a core component of technologies such as autonomous driving and augmented reality. State-of-the-art localization approaches often rely on image retrieval techniques for one of two tasks: (1) provide an approximate pose estimate or (2) determine which parts of the scene are potentially visible in a given query image. It is common practice to use state-of-the-art image retrieval algorithms for these tasks. These algorithms are often trained for the goal of retrieving the same landmark under a large range of viewpoint changes. However, robustness to viewpoint changes is not necessarily desirable in the context of visual localization. This paper focuses on understanding the role of image retrieval for multiple visual localization tasks. We introduce a benchmark setup and compare state-of-the-art retrieval representations on multiple datasets. We show that retrieval performance on classical landmark retrieval/recognition tasks correlates only for some but not all tasks to localization performance. This indicates a need for retrieval approaches specifically designed for localization tasks. Our benchmark and evaluation protocols are available at github.com/naver/kapture-localization.

New Image Retrieval Benchmark for Visual Localization

Poster presentation: International Virtual Conference on 3D Vision

Figure 2 provides an example that illustrates the problem of robustness to viewpoint changes in the context of visual localization. Namely, if we retrieve a set of images for the given query image, the two images in the middle are relevant for landmark retrieval, but not for visual localization. They show the same landmark as the query image, but do not contain overlapping areas, which is crucial for visual localization. Instead, we are looking for images like the two on the right. They show the same landmark and are taken from similar positions.

The main question of this paper is how correlated is landmark retrieval/place recognition performance with visual localization performance?