Response to “Performance characterization in computer vision”