How can I find original youtub video used in vis or vos?And If I want to reproduce your algorithm I have to generate COCO-style annotations by myself,right?

youtubevos / MaskTrackRCNN

MaskTrackRCNN for video instance segmentation based on mmdetection

Apache License 2.0

431 stars 75 forks source link

Open TyroneLi opened 4 years ago

TyroneLi commented 4 years ago

As mention above

linjieyangsc commented 4 years ago

Hi, COCO-style annotations can be downloaded from the VIS challenge website: https://competitions.codalab.org/competitions/20128#participate