Figure 8

Information about the scene available from different layers of the visual pathway. Results are shown over the achromatic contrast and luminance space for two fixed chromatic contrasts: the minimum (zero, on the top) and the maximum in our set (on the bottom). In each case, we assume that sensors at the different representations have equivalent noise level (5% of the signal deviation). This result implies that using sensors of equivalent quality, the cortical representation is more appropriate because it captures more information from the input