In the newshour-chyron gold csv's, the column is scene-label while in the january-slate gold csv's, the column is scene-type. For the purpose of evaluation, these should have the same column title. Since there are plans to also evaluate language identification which does not label scenes, I propose a more general label like label or something similar.
In the newshour-chyron gold csv's, the column is
scene-labelwhile in the january-slate gold csv's, the column isscene-type. For the purpose of evaluation, these should have the same column title. Since there are plans to also evaluate language identification which does not label scenes, I propose a more general label likelabelor something similar.