Closed schmidtp1 closed 3 years ago
e.g. https://github.com/cogsys-tuebingen/mobilestereonet/blob/main/models/MSNet2D.py#L73
@schmidtp1 The 3D convolutions in the 2D model are used for the cost volume computation. We use the naming of 2D or 3D model following the 2D or 3D encoder-decoder.
e.g. https://github.com/cogsys-tuebingen/mobilestereonet/blob/main/models/MSNet2D.py#L73