Equivariant Single View Pose Prediction Via Induced and Restricted Representations