Created a derived Mesh class from TexturedPhotogrammetryMesh named TexturedPhotogrammetryMeshPyTorch3dRendering that handles PyTorch3D rendering. In this derived class, computing the pixel-to-face correspondences is done using PyTorch3D's rasterizer method.
Converting the cameras to PyTorch3D cameras is also handled in this derived mesh class, so PhotogrammetryCamera and PhotogrammetryCamera do not have to deal with PyTorch3D. I created a get_instrinsic_camera_properties() method in the PhotogrammetryCamera class so that the derived mesh using PyTorch3D for the pix2face computation has access to the cameras to convert them to PyTorch3D cameras.
Created a derived Mesh class from
TexturedPhotogrammetryMeshnamedTexturedPhotogrammetryMeshPyTorch3dRenderingthat handles PyTorch3D rendering. In this derived class, computing the pixel-to-face correspondences is done using PyTorch3D'srasterizermethod.Converting the cameras to PyTorch3D cameras is also handled in this derived mesh class, so
PhotogrammetryCameraandPhotogrammetryCamerado not have to deal with PyTorch3D. I created aget_instrinsic_camera_properties()method in thePhotogrammetryCameraclass so that the derived mesh using PyTorch3D for the pix2face computation has access to the cameras to convert them to PyTorch3D cameras.