create_labels() returning empty variables
OneManArmy93 opened this issue · comments
I tired to execute the this funtion def create_labels(places, size, corners, resolution = 0.5, x=(0, 90), y=(-50, 50), z=(-4.5, 5.5), scale=4, min_value=np.array([0., -50., -4.5])):
but the returned variables sphere_center, train_corners
have the size of (0,3) and (0,8,3)
but they are empty.
Any idea about that?
Thank you
May be there are no objects in the ground truth. Try with the frame which has some cars in it
Yes im working with a frame where there is a car.
I managed to find the root of the problem, it is in because the z_logical
is alwayse False
hence xyz_logical
is always False
so sphere_center, train_corners
are always empty. I noticed the same behavior in center_to_sphere(places, size, resolution=0.50, min_value=np.array([0., -50., -4.5]), scale=4, x=(0, 90), y=(-50, 50), z=(-4.5, 5.5))
because places[:, 2] < z[1]
this is always False
Can you explain the reason of using x_logical, y_logical, z_logical
and why those values, because they are genetaring the issue. Thx
EDIT: I changed the z values to z=(-4.5, 20)
and now it is not returning empty variables, but does that change affects anything?
z is the vertical direction of point cloud, you can set it as z=(-3, 1) for kitti data set, actually it is filtering out the ground truth objects which are out side the range of experiment range.
ok, i'll try to do that but i set it at (-4.5, 20)
and returned not empty viarbles. I guess i'll see how it goes from there.
Regarding my question I post it on an other issue:
what does this function dor create_objectness_label
?
why there is 3 loss functions?
and I couldn't find the declation of the cordinate
function and What does it do?
Thank you