On Measuring* the Ecological Validity of Local Figure-Ground Cues