pytorch3d
pytorch3d copied to clipboard
Problems with rendering
Can this work be used to render BFM face models? I get this warning when I try to render this model:Bin size was too small in the coarse rasterization phase. This caused an overflow, meaning output may be incomplete. To solve, try increasing max_faces_per_bin / max_points_per_bin, decreasing bin_size, or setting bin_size to 0 to use the naive rasterization. Later I set bin_size to 0, but I get the rendered video result is completely blank, am I doing something wrong?
Can you use the plotly_vis functionality to make a 3d plot of the camera and the object together, to check they are lined up roughly right. Also, you could look at the fragments from rasterization - if they aren't empty then your problem of "blank" output would have something to do with shading or something later, which can be debugged.
I also encountered a blank rendering. Later, I found that it was because I didn't set the focal_length and principal_point; I just used the default value of PerspectiveCameras, especially when the in_ndc mode was turned off. After correctly setting these values, the rendering became normal. Noted it here in case anyone encounters the same problem.