How real is real enough? optimal reality sampling for fast recognition of mobile imagery

We present the first study to discover optimal reality sampling for mobile imagery. In particular, we identify the minimum information required for fast recognition of images of directly perceivable real-world buildings displayed on a mobile device. Resolution, image size, and JPEG compression of images of façades were manipulated in a same--different recognition task carried out in the field. Best-effort performance is shown to be reachable with significantly lower detail granularity than previously thought. For best user performance, we recommend presenting images as large as possible on the screen and decreasing resolution accordingly.