This workshop is to foster research on image retrieval and landmark recognition by introducing a novel large-scale dataset, together with evaluation protocols.

Call for Participation to the Challenge

Large-Scale Landmark Recognition: A Challenge
https://landmarkscvprw18.github.io/


We invite participation in the Google Landmark Recognition and Retrieval Challenges hosted by Large-Scale Landmark Recognition: A Challenge (Landmarks) workshop in conjunction with CVPR’18 at Salt Lake City, UT, USA. The goal of these challenges is to foster research in large-scale landmark recognition and image retrieval. Today, a great obstacle to research in these fields is the lack of large annotated datasets. We present the largest worldwide dataset for these tasks to date. The recognition track challenges participants to build models that recognize the correct landmark in a dataset of challenging test images. In the retrieval track, participants are expected to retrieve images containing the same landmark. The top-ranking teams will be invited to present their methods at Landmarks workshop at CVPR’18.

The challenges are hosted on Kaggle:

Recognition track:
https://www.kaggle.com/c/landmark-recognition-challenge

Retrieval track:
https://www.kaggle.com/c/landmark-retrieval-challenge

Prize:
For both challenges, Google supports cash prize!
- 1st place: $1,250
- 2nd place: $750
- 3rd place: $500

Timeline:
- Challenge Launch: 2/2/2018
- Challenge Deadline: 5/22/2018
- Workshop: held with CVPR’18

Organizers:
Bohyung Han (POSTECH; bhhan@postech.ac.kr), Andre Araujo (Google; andrearaujo@google.com), Shih-Fu Chang (Columbia University), Ondrej Chum (Czech Technical University), Torsten Sattler (ETHZ), Jack Sim (Google), Giorgos Tolias (Czech Technical University), Tobias Weyand (Google), Xu Zhang (Columbia University)