Hi,
I found that some bounding boxes in the dataset appear to exceed the corresponding image size if they are interpreted as pixel coordinates.
For example, one sample has:
If the bbox is interpreted as pixel coordinates, the x ccoordinates are outside the image width:
width = 543
x1 = 622
x2 = 656
It is the wrong bounding box!
Could you please clarify the coordinate system used for the ... annotation?
Thanks
Hi,
I found that some bounding boxes in the dataset appear to exceed the corresponding image size if they are interpreted as pixel coordinates.
For example, one sample has:
If the bbox is interpreted as pixel coordinates, the x ccoordinates are outside the image width:
width = 543
x1 = 622
x2 = 656
It is the wrong bounding box!
Could you please clarify the coordinate system used for the ... annotation?
Thanks