24/7 Place Recognition by View Synthesis

Example query images from the newly collected 24/7 Tokyo dataset. Each place in the query set is captured at different times of day: (a) daytime, (b) sunset, and (c) night. For comparison, the database street-view image at a close-by position is shown in (d). Note the major changes in appearance (illumination changes in the scene) between the database image (d) and the query images (a,b,c).
Authors: Akihiko Torii, Relja Arandjelović, Josef Sivic, Masatoshi Okutomi, Tomas Pajdla
Citation BibTeX |
Code: Code (dense VLAD description, feature matching, view synthesis) (1.1GB). (ver01 08/06/2015)
Data:
Query images (4.1GB). (ver 01 08/06/2015)
Query images (4.1GB). (ver 02 20/06/2015) (UTM added in csv files)
Query images (4.1GB). (ver 03 21/09/2015) (GPS, UTM of 0757-0765 were corrected. These images did not appear in our CVPR 2015 paper.)
Query images - subset used in our CVPR 2015 paper - (1.1GB).