In train.py file,
if propagated_mask.sum() > 100:
gaussians.densify_from_depth_propagation(viewpoint_cam, propagated_depth, propagated_mask.to(torch.bool), gt_image)
the "gt_image" is not mentioned before. This means that each time this operation is performed, the used "gt_image" is from the previous iteration?
In train.py file,
if propagated_mask.sum() > 100:
gaussians.densify_from_depth_propagation(viewpoint_cam, propagated_depth, propagated_mask.to(torch.bool), gt_image)
the "gt_image" is not mentioned before. This means that each time this operation is performed, the used "gt_image" is from the previous iteration?