Illuminating research on the perception of shape: more than just lighting up candle wax

Have you ever wondered why statues made from wax or quartz look different to ones made from stone? It could have something to do with the perception of shape and opacity.

Scientists based in the School of Optometry and Vision Science at UNSW recently conducted a research study looking at how our brain perceives the 3D shape of opaque and translucent objects. They found that the perception of bumps on the surface of the same 3D object look less deep and less steep than identical bumps on an opaque object.

The work was innovative, because the researchers were not limited to using graphical simulations; they chiselled out crevices in an actual candle, and then photographed it. They then painted it using an opaque acrylic paint and rephotographed for comparison. Not only did it look more opaque, the surface also looked bumpier.

The researchers went on to show that the perception of curvature in similar translucent objects was highly underestimated when compared with opaque objects and also ground truth (the objects actual 3D geometry). The effect of opacity on the perception of shape was explained by differences in the steepness of the shading gradients in the image.

“This isn’t just about contrast, because if we match image contrast we still see the difference in perceived shape,” Dr Juno Kim stated, one of the authors on the published work.

What appears to be important for the perceptual difference is the ability of light to penetrate through the surface of a non-opaque object. This essentially allows light to get to surface regions on the far side of the object that wouldn't normally receive direct illumination if they were completely opaque.

“This highlights the importance for vision researchers to continue looking for image cues that seem to tell us how the world is structured,” Dr Kim said.

The study was led by Nahian Chowdhury a research student supervised by Dr Kim (UNSW) and in collaboration with Dr Phillip Marlow (University of Sydney). For more information about this research, the original article is published for open access through the Journal of Vision website: