Dataset Hub

NYC‑Indoor‑VPR is a unique indoor visual place recognition dataset created by New York University, comprising over 36,000 images captured from 13 crowded scenes across New York City under varying lighting and appearance conditions. The dataset was built using a semi‑automatic labeling method to establish ground truth locations for each image. It is primarily intended for indoor localization and navigation research, especially for robots and assistive navigation systems, addressing the challenges of visual aliasing and occlusion in indoor environments.

GSV‑Cities is a large‑scale visual place‑recognition dataset containing approximately 530 k images from over 62 k distinct locations worldwide. Each location is represented by at least 4 and up to 20 images, with a minimum physical separation of 100 m between locations.

The Nordland dataset captures images from a 728‑km railway journey in Norway, divided into the four seasons: spring, summer, autumn, and winter. The dataset is organized into four folders named after the seasons, each containing 35,768 images. Images correspond one‑to‑one across folders. For each traverse, corresponding ground‑truth data are provided in the specified .csv files. Additionally, the dataset includes a file named `nordland_imageNames.txt` that provides a filtered list of images, excluding segments captured when the train speed was below 15 km/h.