SIREN showcases an example where they use the gradient of the SDF during training. The gradient is the spatial partial derivative of the SDF, closely related to the normal vector.
This PR adds the ability to compute the gradient along with the SDF samples. This solution sould be simple to expand to voxel models.
SIREN showcases an example where they use the gradient of the SDF during training. The gradient is the spatial partial derivative of the SDF, closely related to the normal vector.
This PR adds the ability to compute the gradient along with the SDF samples. This solution sould be simple to expand to voxel models.
Example: