Is Image-based Object Pose Estimation Ready to Support Grasping?