Welcome to visuallocalization.net!

In order to evaluate visual localization and place recognition performance over longer periods of time, we provide benchmark datasets aimed at evaluating 6 degree-of-freedom pose estimation accuracy over large appearance variations caused by changes in lighting conditions, seasons, snow, and day-night changes. Each dataset consists of a set of reference images, together with their corresponding ground truth poses. A triangulated 3D model is also be provided for each dataset. For detailed information about the datasets, please see our CVPR 2018 paper.

UPDATE 2018-10-05

We have now made available two of the three datasets, the RobotCar-Seasons dataset and the CMU-Seasons dataset. We are currently working on getting the Aachen Day-Night dataset up online as soon as possible.

We are also working on setting up a benchmarking server for automatic evaluation of your localization results. If you are working on a publication and would like to evaluate your method, please see the instructions for evaluation in the readme for the corresponding dataset.

The datasets

Aachen Day-Night dataset (to-be released)

The Aachen Day-Night dataset consists of 4,328 database images for map-building, and 922 query images for evaluation, taken from a smart phone in an urban environment. The dataset primarily focuses on evaluating day-to-night visual localization.

CMU-Seasons dataset (released)

The images of the CMU-Seasons dataset are a subset of the CMU Visual localization dataset. The dataset contains 7,159 images for map building, and 75,335 query images for evaluation. Like the RobotCar-Seasons dataset, the images are taken from a car, and the dataset is aimed towards an autonomous driving scenario. The dataset contains urban and suburban areas, as well as country road scenes with very few man-made structures and large amounts of vegetation.


RobotCar-Seasons dataset (released)

The RobotCar-Seasons dataset consists of a subset of the Oxford RobotCar dataset, and contains 20,862 for map-building and 11,934 query images for evaluation. The query images contain a large variety of different conditions, including snow, rain, night-time images. The images are taken from a car and aimed at an autonomous driving scenario.



Once you submit your address you will be forwarded to a confirmation page