It means that a larger image field can be sampled from a lower resolution copy without much loss in comparative data, only the number of data points to be manipulated.