Extracted Features of Amsterdam Cross-time Dataset

Datacite citation style:
Yildiz, Burak; Khademi, Seyran (2021): Extracted Features of Amsterdam Cross-time Dataset. Version 1. 4TU.ResearchData. dataset. https://doi.org/10.4121/14572644.v1
Other citation styles (APA, Harvard, MLA, Vancouver, Chicago, IEEE) available at Datacite

Dataset

This collection contains feature files where the features are extracted on Amsterdam Cross-time dataset using various methods and models such as SIFT, LIFT, and pre-trained VGG-16, ResNets, NetVLAD, AP-GeM and supervisely and self-supervisely trained models. The details of feature extraction procedure and other details can be found on https://github.com/seyrankhademi/cross-time-dataset.

History

  • 2021-05-12 first online, published, posted

Publisher

4TU.ResearchData

Format

Pickle file

Organizations

TU Delft, Faculty of Electrical Engineering, Mathematics and Computer Science Intelligent Systems

DATA

Files (11)