You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
dim is the dimension (size) of the 3D box, noc is the normalized object coordinate. So noc * dim is to recover the absolute scale of the 3d coordinates.
I also want to know the meaning of 'w2d'。 i think is the confidence of 2d-3d correspondences. But why the dimension of w2d is [bs, num_point ,2 ] ,why the last dim is 2?
感谢您精彩的工作,最近在学习您提供的EPro-PnP-6DoF的代码时,遇到了一些困惑希望得到您的解答。
在train.py文件中,dim 代表的是什么?是每张图片中目标物体的中心点在相机下的x, y, z坐标吗?
noc 具体指的是什么意思呢?我观察到noc 来自与 resnet_rot_head.py中的返回值x3d, 那么在train.py中, x3d = noc * dim[..., None, None]又代表什么意思呢?
The text was updated successfully, but these errors were encountered: