Official Repo for Ground-aware Monocular 3D Object Detection for Autonomous Driving / YOLOStereo3D: A Step Back to 2D for Efficient Stereo 3D Detection
This is a rather important bug found in the code. Thank you for reporting.
I was rather lucky because (EL - 0.5*h_mean) is numerically reasonably close to 1.0.
Is there a parenthesis missing in the code.
y_shifts_base = F.relu( h_mean * (yy_grid - cy) / 2 * (self.relative_elevation - 0.5 * h_mean) ) / (yy_grid.shape[1] * 0.5) # [1, H, W]
https://github.com/Owen-Liuyuxuan/visualDet3D/blob/master/visualDet3D/networks/lib/look_ground.py