The videos are broken down in chunks of 1000 frames per chunk.
The fields below should be copied into CVAT when creating a new task.
Name: ConferenceFolder-FolderID-VideoName #omit .mp4 (e.g. "ParkCity2019-170211-ch1_video_01_frame_00022001_frame_00023000")
Labels: tool ~radio=type:__undefined__,grasper,cottonoid,muscle,suction,drill,string
frame ~radio=type:start,stop
Bug Tracker:
Dataset Repository: https://github.com/guillaumekugener/surgical-training-project.git
Source: Local
Z-Order: n/a
Overlap Size: n/a
Segment Size: 5000 #is this default? capped? what does this represent?
Start Frame: n/a
Stop Frame: n/a
Frame Filter: n/a