Home

Announcements

• A draft version of the task overview paper is available here. In your working notes paper you can refer to this paper and don't have to repeat the information about dataset, task, evaluation. Instead focus on your proposed methods.
• The test-set for the DIRSM and FDSI subtask is released. Please note, that we will release an additional test-set for the FDSI-subtask in the next couple of days.
• We changed the submission for the DIRSM task. Required run 2 should be based on metadata-information only and not visual only (as it is already covered in required run 1).

Task steps

Participation in this task involves the following steps:

• Design your system on the development data (devset): download the development data and design your approach for the two subtasks;
• Run your system on the test data (testset): when the test data is released, run your system on the test data. You may submit up to 5 runs;
• Receive your evaluation results from the task organizers. Your results will be reported in terms of the official evaluation metric of the task, which you must report in the working notes paper. However, you are encouraged to also report your results with the evaluation metrics of your choice;
• Write and submit your working notes paper. Make sure that your working notes paper describes exactly the conditions for which you submitted runs in Step 2 above.

Task schedule

May 1, 2017	Development data release
June 1, 2017	Test data release
August 17, 2017	Run submission due
August 21, 2017	Results returned to the participants
August 28, 2017	Working notes paper initial submission deadline
August 30, 2017	Working notes review returned to the participants
September 4, 2017	Camera ready working notes paper due
September 13-15, 2017	MediaEval Workshop, Dublin, Ireland

Task organizers

• Benjamin Bischke, German Research Center for Artificial Intelligence (DFKI), Germany (first.last at dfki.de)
• Damian Borth, German Research Center for Artificial Intelligence (DFKI), Germany (first.last at dfki.de)
• Christian Schulze, German Research Center for Artificial Intelligence (DFKI), Germany (first.last at dfki.de)
• Alan Woodley, Queensland University of Technology (QUT), Australia
• Venkat Srinivasan, Virginia Tech (Blacksburg VA), US