We would need some code to deal with deleted track.
The only way to delete false positive in bulk is by deleting the track in grid, but this will leave the bounding_box in the annotation but with empty track. For example:
04.09.2023-04-Right_RE_test_frame_00000184.png,15259591,"{""clip"":123}",79,0,"{""name"":""rect"",""x"":3899.8646916677476,""y"":916.2452700052339,""width"":50,""height"":36}","{""track"":""""}"
Also ensure that bounding boxes IDs are unique per frame (because we copy+paste annotations across frames this is likely to happen)
We would need some code to deal with deleted track. The only way to delete false positive in bulk is by deleting the track in grid, but this will leave the bounding_box in the annotation but with empty track. For example:
04.09.2023-04-Right_RE_test_frame_00000184.png,15259591,"{""clip"":123}",79,0,"{""name"":""rect"",""x"":3899.8646916677476,""y"":916.2452700052339,""width"":50,""height"":36}","{""track"":""""}"
Also ensure that bounding boxes IDs are unique per frame (because we copy+paste annotations across frames this is likely to happen)