Counter shaded animal patterns: from photons to form

Many different kinds of animals have camouflaged bodies. In many cases the pattern of colouration features a darker skin, or fur, on the surface of the body that is closer to the sun, and lighter shading on the other side. For example, fish are often dark along their backs and light along their bellies. The same is true for many other animals for example deer, birds, lizards and many insects. This pattern of colouration is known as 'counter-shading'. In this project we aim to understand how counter-shading might provide a useful source of camouflage for animals, and why it has evolved. Does it make them harder to detect, or is their apparent shape, as perceived by a potential predator, changed by this form of colouration? Even a small advantage in being more difficult to detect would enhance the animal's chances of survival, and increase the liklihood of its passing that colouration advantage on to the next generation. There are two ways in which this pattern of colouration might have evolved in order to make an animal harder to detect. First, the animal may simply be trying to match the background. When viewed from above the animals back is darker and matches the ground, and when viewed from below its lighter underbelly matches the sky, so the camouflage could be simply an attempt at background matching. Second, the patterning could represent an attempt to minimise the shading across its surface, so that it appears flatter than it actually is. We know from research in human vision that the shading on the surface of an object helps us to perceive the 3-dimensional (3D) shape of the object (called shape-from-shading). This explains how we can perceive 3D objects in black and white photographs, even though the photograph is actually just a flat surface covered with varying amounts of black ink. So, counter-shading may have evolved in order to confuse these shape-from-shading processes in the brain. A perceptually 'flatter' prey animal may be more difficult to see, or less desirable to eat. The first stage of our project will be to examine counter-shading on prey animals in detail. On one has ever measured the exact patterning on animals and tested to see if the patterns found match the pattern that would be expected for each of the explanations given above. We will measure the 3D shape of individual animals, and their counter-shading. Using the measurements, we will create 3D computer models of animals and their shading. In theoretical studies, we will develop mathematical models that predict what patterns of counter-shading would be ideal for hiding the animal. We will then be able to test these predictions by comparing with the physical measurements. Our computer simulations will show us what patterns of shading are most helpful in hiding an animal. In the second stage of the project, we will test whether the 'most helpful' patterns are actually harder to detect. We will use humans and birds as observers in perceptual experiments that test whether the best shading patterns allow a prey animal to remain hidden for longer. We start by studying humans, because a great deal is already known about shape-from-shading in the human visual system. We also test birds, because their brains are organised rather differently, yet given the frequency of counter-shading in prey, one would predict that non-mammalian predators should also be fooled by this form of camouflage. Finally, we can take what we learn from these simulations and test the results in the real world. By attaching treats to printed cardboard tubes that are distributed around a real outdoor environment, we can see how quickly the tubes are found by wild birds. If the tubes are taken less quickly then we can assume that the shading is much better camouflage. At the end of our project we expect that we will be much closer to an explanation of why counter-shading has evolved in many types of animal.

Many species are counter-shaded: the dorsal surface is darker than the ventral. It has been proposed that counter-shading offers the animal camouflage. There are two potential accounts of its evolution: (i) counter-shading enables the animal to match its background: when viewed from above the dorsal surface is darker and matches the ground, when viewed from below its lighter ventral matches the sky. (ii) Counter-shading is self-shadow concealment. The image of a 3D uniform coloured object will exhibit a shading pattern, determined by its shape and the light-source direction. Visual shape-from-shading brain processes allow humans to perceive 3D shape even though the retinal image is 2D. A counter-shaded animal disrupts the pattern of shading coming from the light-shape interaction. In the extreme, the shading could cancel out, impeding detection of 3D shape and affecting visibility. Using calibrated cameras we will quantify counter-shading in animals and develop mathematical models to test whether the observed patterns match those expected for the hypotheses above. The models will predict which patterns of shading are best suited to hiding the animal. In laboratory experiments, we will test how well optimal counter-shading patterns fool human visual systems, to probe the details of counter-shading processing. We will also test using bird visual systems, to examine the generality of success of the counter-shading strategies. Finally, field studies will provide the ultimate test: whether the optimal shading patterns from our simulations do improve concealment from birds in natural lighting environments. Our project will determine why counter-shading has evolved in diverse species and the extent to which mammalian (human) and bird visual systems possess visual mechanisms sensitive enough to detect prey despite theoretically optimal counter-shading.

Companies/public sector bodies [a] This work has potential commercial application, although these lie downstream of the current application. Shape from shading provides low-level cues to structure for simple brains, consequently our findings should be of interest to those involved in developing artificial vision systems. Further, understanding how shading can hinder detection may allow use of shading to enhance detection for reasons of safety (overhead power lines viewed from aircraft) or information transfer (effective design of roadside signage). [b] One of the first aims of the grant is the extension of a stereoscopic photogrammetry system to utilise calibrated cameras - enabling the recovery of accurate spatial and chromatic information. This system would enable the accurate reconstruction of 3 dimensional scenes, giving accurate colour and distance information. Such a system would be very useful in the human factors industry, for example facilitating the assessment of the visibility of warning systems. Cameras calibrated by Dr Lovell are already used by Railtrack to assess railway sign and signal visibility. [c] A further stage in the development of the photogrammetric scanning system is the calibrated recovery of surface reflectance and the estimation of the BRDF function. Such a system would be very useful to those in the game development and film special effects industries attempting to render scenes that appear realistic. Methods and activities We plan to establish a website sharing the 3D scans made of various countershaded animals. These scans would be useful to academics interested in further examining pigmentation patterns and camouflage. The existence and utility of the site will be advertised to the community via our conference presentations, journal publications, and by posting to relevant email lists and user groups. The Hunterian Museum in Glasgow has a display linking historical camouflage work by Sir Graham Kerr and Hugh Cott to more recent work by Ruxton. Ruxton has closely links with the museum, as honorary curator of zoology, and will keep this display updated to reflect the latest developments of this project. The Applied Vision Association (AVA) provides a useful platform for the presentation our findings to academics in related fields and to industrial partners. We will budget for two people to attend one AVA meeting each (Xmas or Annual Easter) near the end of the project.