Can I use pytorch 3d to calculate the gt, predict of the vertices of a 2d image with Differentiable rendering loss?
@nobug-code I'm not sure I understand what you are asking. Do you mean predict 3D vertices of a mesh from a 2D image?
Can I render if I have a 2D image and given coordinates?
Do you have the 3D vertex coordinates of the object? What are you trying to render? What do you want as the output? Please provide some more details about what you are trying to achieve.
@nobug-code please provide more details with examples if you still need help with this issue.
I'm sorry for the late contact. What I'm trying to implement is the neural-render-pytorch version of the feature, so I'm going to use it.
I was trying to render a 3d z axis of 1 from a point.
thanks for your answer!
@nobug-code can you provide pseudo code or more detailed example of exactly what you are trying to do?
What I want to do is to draw 3d renders using 2d points information (fix z axis to 1)
If what you mean here is take the points from an image and assign them all a z value of 1, and then render the resulting mesh (which is flat) in PyTorch3d, then that is demonstrated in the attached, which I run from the tutorials folder.
@nobug-code if this answers your question please close this issue.
Most helpful comment
If what you mean here is take the points from an image and assign them all a z value of 1, and then render the resulting mesh (which is flat) in PyTorch3d, then that is demonstrated in the attached, which I run from the tutorials folder.
render_logo.txt