Challenge details 2019

Details on inputs/outputs

Predicting local character IDs means that it is not required to recognize each character globally (in an entire movie), but locally (within a set of 5 clips). Practically, this means that the submissions should predict unique character IDs that are consistent within a given set of 5 clips, that is perform a local character re-identification. Note, that while we also provide annotations with global character IDs for completeness, it is not required to generate such global IDs, but only to predict consistent IDs within each set of 5 clips (as demonstrated by the provided local IDs). The segmentation of annotations into sets of 5 clips is performed sequentially and is also provided for participants' convenience. All of the above can be accessed here.

1. An example input/output for the 1st task:

2. An example input/output for the 2nd task:

3. An example input/output for the 3rd task:

Note: In our automatic evaluation of the 3rd task we will focus on sentence content but ignore the predicted IDs. It is therefore required that the participants also submit their results for the 2nd task. (We are also considering an additional human evaluation to assess the correctness of the predicted IDs.)

Submission servers

Task 1 "Multi-Sentence Description": https://competitions.codalab.org/competitions/20669

Task 2 "Fill-in the Characters": https://competitions.codalab.org/competitions/20668

Task 3 "Multi-Sentence Description with Characters": https://competitions.codalab.org/competitions/21051

Contact

lsmdc2015+19 at gmail.com