I'm wondering how the model could draw the bounding box on the input image.
The image was first resize to 224x224, then after the convolutions, 14x14 feature map could be abtained.
The bounding box's coordinates range from [0,13] rather than the [0,223].
How could I obtain a result with the bounding box drawn in it?
I've tried interpolation='spline16' but it didn't work well.
Thanks!!!!~~~ :)