tapnet
tapnet copied to clipboard
Understading about annotation
You mention The points (float32 tensor with 3 axes; the first is point id, the second is time, and the third is x/y).
What is x/y here? I read the pkl file, and the value looks like [0.41469392, 0.53467155] so I guess, you are normalized it.
x = position in horizon/width y = position in vertical / height
Am I correct? Thanks Dat
Yes, what's stored in the pickle file is normalized. As usual for image coordinates, 0,0 is the upper-left corner.
The reader re-normalizes these to actual pixel coordinates, i.e., multiplies by [width, height].