Dataset split & Query portrait pictures & resolution of key frames on Character Identification task.

[ZhengPeng7]

Thanks for your great work! But I have several questions that need your help.

They are split into train, val, test set with 2088 identities with 639.9K instances, 821 identities with 336.6K instances and 876 identities with 364.2K instances respectively.

-- Could you provide the PIDs for the test? (I'm using set(pid_testset) - set(pid_trainset) as the test pid.)

we download a portrait from homepage of each credited cast, which will serve as the query portraits for the character identification tasks

-- Could you provide the link to download the protraits you collected, which were used in your experiments?

Thanks a lot.

[ycxioooong]

Hi @ZhengPeng7, we will make accordingly revisions.

[ZhengPeng7]

Thanks for your quick reply👍.

[ZhengPeng7]

Besides, could you provide the keyframes with 720P as is stated in the original paper? Because all the annotations are made in 720P.... If trained on 240P, details of the facial information that helps the identification may be lost.