Repository navigation
Why z0 is set to be 50? #36
Description
Hi, thanks for your excellent work! I have two questions.
- According to my understanding, z0 represents the height of the ceiling from the camera plane，isn't it 1.6 meters? why z0 is set to be 50?
line 95 inference.py
# Init floor/ceil plane
z0 = 50
_, z1 = post_proc.np_refine_by_fix_z(*y_bon_, z0)
- Does variable tol mean the threshold of signal when recovering all planes? tol is 0.05 in paper but abs(0.16 * z1 / 1.6) in code.
line 106 inference.py
cor, xy_cor = post_proc.gen_ww(xs_, y_bon_[0], z0, tol=abs(0.16 * z1 / 1.6), force_cuboid=force_cuboid)
line 81 post_proc.py
invalid = (n < len(vec) * 0.4) | (l > tol)
Activity
sunset1995 commentedon Aug 19, 2021
Hi, thanks for trying our work.
Re 1:
The 3D results is up to a scale (i.e., we can scale all 3D corner by a constant factor and the results on the 2D image is still the same), so we are free to chose any z0>0 during computing.
Re 2:
Yes. The
tolis used to determine the inliers given a 1D signal (the x or y coordinate of a wall on the floorplane).marcomiglionico94 commentedon Feb 17, 2022
Hello I have a question about the coordinates
uvpresent in the json file. I would like to ask how do you converted those coordinates into a range of [0,1]. I am trying other networks and then try to visualize them using the layout visualizer code. The main issue is that other networks output the results in a range that is not [0,1]. For example:[[ 1.7968756 -3.6718748 ]
[ 1.1718756 -3.6718748 ]
[ 1.1328129 -1.9921873 ]
[-2.3046873 -1.9921879 ]]
When I try to use the layout visualizer code with these coordinates it doesn't work. How can i convert those into the same [0,1] range?