We can visually estimate surface properties of natural objects, not only colour and lightness, but also glossiness, translucency, and softness. How does the human brain accomplish this job? Considering the fact that the image of a surface is a result of highly complex optical process including reflection, refraction, and scattering, one may think that the estimation of surface properties should involve deep, complex neural computations including 3D shape reconstruction. However, psychophysical evidence suggests that the perception of some surface qualities uses shallow and simple computations based on simple statistics or features in the 2D image. For example, Beck  have pointed out that the presence of highlights in the surface image is a strong cue for the perceived glossiness. Our recent analysis demonstrated that the perceived lightness and glossiness of a natural surface strongly depend on skew in the luminance histogram of the image [2, 3]. Is this also true for the other surface properties such as translucency and metallicity? In optics, translucent objects generally have strong scatterings of the light inside the material, and metallic ones a larger amount of mirror reflections than diffuse reflections. As a result, these materials exhibit specific patterns on the surface image distinct from those of the other materials in complicated ways. It seems difficult to describe key image features for those properties. However, we here demonstrate that a simple remapping of the image luminance can dramatically alter the perceived material.