City-scale landmark identification on mobile devices

摘要: With recent advances in mobile computing, the demand for visual localization or landmark identification on devices is gaining interest. We advance state of art this area by fusing two popular representations street-level image data — facade-aligned and viewpoint-aligned show that they contain complementary information can be exploited to significantly improve recall rates city scale. also feature detection low contrast parts data, discuss how incorporate priors a user's position (e.g. given noisy GPS readings network cells), which previous approaches often ignore. Finally, maybe most importantly, we present our results according carefully designed, repeatable evaluation scheme make publicly available set 1.7 million images with ground truth labels, geotags, calibration as well difficult cell phone query images. provide these resources benchmark facilitate further research area.