PHASE 2 | Facebook AI Image Similarity Challenge: Descriptor Track
Advance the science of image similarity detection, with applications in areas including content tracing, copyright infringement and misinformation. In the "Descriptor Track" participants generate image descriptors that enable the problem to be solved at scale. Competition hosted by Facebook. #society
For ongoing research and development, you can still access the DISC21 dataset and other information about the competition here:
Welcome to Phase 2
Welcome to Phase 2 of the Facebook Image Similarity Challenge! Congratulations for all the work you've already done, and good luck in this final stage.
How to access the Phase 2 data
For Phase 2, you only need to download the new set of 50,000 query images. The reference images will be the same ones you used in Phase 1.
As with Phase 1, you will be able to choose to access the data from one of 3 AWS regions:
- US West (Oregon)
- Asia (Singapore)
- Europe (Frankfurt)
As you already know, you can download the data with AWS CLI using a command like below. Note that you may or may not need the
--no-sign-request flag depending on whether you have an AWS profile configured.
aws s3 cp s3://<<SEE DATA DOWNLOAD PAGE FOR LOCATION>> ./ --recursive --no-sign-request
Tools like s5cmd may also be able to provide faster performance than AWS CLI.
Make your submission(s)
Once you've downloaded the Phase 2 data, you can read it in with your existing model and generate a submission file just as you did in Phase 1. To make your submission, click on the "Submissions" link and submit your file as with Phase 1.
Note that aside from reading in the new data, the code and model (including model weights) generating your submission must be identical to what was submitted in your "End Phase 1 Submission". For the Descriptor Track, the descriptors in your submission should be identical to what was submitted in your "End Phase 1 Submission".
Please also review and make sure you are remaining in compliance with the Rules on Data Use. The following rules are particularly relevant to Phase 2:
- Adding annotations to any images used in Phase 2 is prohibited.
- You must treat each new image of the reference set independently and without any interaction with other reference images.
You may make more than one submission if you made an "End Phase 1 Submission" for more than one model, or if you believe a previous Phase 2 submission was invalid. You are limited to no more than 3 submissions in total during Phase 2. In most cases, if you are using a single model, you will only need to make one submission in Phase 2. Your highest ranking submission will determine your place on the final leaderboard.
Please keep the following in mind as you prepare your submission and await final scoring.
- Make your submission as early as possible to avoid bottlenecks during busy periods which might result in your submission being queued.
- Be patient waiting for your score after submission. This can take a while, particularly for Descriptor Track scoring, since it requires a similarity search across 50 billion image pairs.
- Rest assured that if you are able to successfully upload a valid file, it will get scored.
Oct. 27, 2021, 11:59 p.m. UTC
Competition End Date
Predicted matching scores for pairs of query and reference images are evaluated using micro-average precision.
Predicted vector representations of images (up to 256 dimensions) are compared with Euclidean vector distance to generate matching scores as in the Matching Track, which are then evaluated using micro-average precision.
Note: Prizes will be awarded to the teams that will share their solutions under an open source license per the Competition Rules. However, willingness to share solutions will not impact the leaderboard positions, so that top performers will still maintain their leaderboard rankings regardless of whether they receive a cash prize.
Prize generously supplied by Facebook AI.
The challenge is supported by Pinterest, BBC, Getty Images, iStock and Shutterstock.
NO PURCHASE NECESSARY TO ENTER/WIN. A PURCHASE WILL NOT INCREASE YOUR CHANCES OF WINNING. The Competition consists of two (2) Phases, with winners determined based upon Submissions using the Phase II dataset. The start and end dates and times for each Phase will be set forth on this Competition Website. Open to legal residents of the Territory, 18+ & age of majority. "Territory" means any country, state, or province where the laws of the US or local law do not prohibit participating or receiving a prize in the Challenge and excludes any area or country designated by the United States Treasury's Office of Foreign Assets Control (e.g. Cuba, Sudan, Crimea, Iran, North Korea, Syria, Venezuela). Any Participant use of External Data must be pursuant to a valid license. Void outside the Territory and where prohibited by law. Participation subject to official Competition Rules. Prizes: $50,000 USD (1st), $30,000 (2nd), $20,000 USD (3rd) for each of two tracks. See Official Rules and Competition Website for submission requirements, evaluation metrics and full details. Sponsor: Facebook, Inc., 1 Hacker Way, Menlo Park, CA 94025 USA.