KeywordsHuman Observer Depth Perception Nonhuman Animal Motion Parallax Occlude Object
Occlusion is a monocular depth cue produced by partially overlapping objects: Objects that partially block other parts of the scene are perceived to be closer to an observer than the blocked objects.
Like motion parallax, occlusion is a monocular depth cue that does not require integrating information from two retinas. Unlike motion parallax, however, occlusion is a pictorial depth cue that is available in static images. In addition to using occlusion for ordering objects in depth, human observers have a strong tendency to perceive partially occluded objects as being completed behind an occluded surface, a process called visual completion or amodal completion. However, the presence of amodal completion in non-primate species is less explored and, in some cases, controversial.
Occlusion and Visual Completion in Humans and Nonhuman Animals
In humans, amodal completion is normally demonstrated by using two-dimensional images similar to those shown in Fig. 1. The observers are asked, either directly or indirectly, to judge whether the occluded image (e.g., a rectangle, Fig. 1a) is more similar to the completed version of that image (e.g., a full rectangle, Fig. 1b) or an incomplete version that presents the portions not obscured by the occluded (Fig. 1c). Although this approach has successfully been used for a number of mammalian species (reviewed by Fujita 2006), many other efforts to demonstrate amodal completion using such images were either unsuccessful (pigeons: Sekuler et al. 1996) or required additional contextual cues or experimental manipulations to induce amodal completion (baboons, Deruelle et al. 2000; pigeons, Nagasaka et al. 2007).
In contrast, experiments using real three-dimensional objects consistently revealed amodal completion in a variety of species (fish, Sovrano and Bisazza 2008; baboons, Deruelle et al. 2000). Moreover, pigeons have recently been shown to use occlusion for ordering two-dimensional objects in depth (Cavoto and Cook 2006), further suggesting that they may be capable of amodally completing objects in certain circumstances.
On the other hand, pigeons have also been shown to experience difficulties in recognizing pictures of occluded objects without additional experimental manipulations (Aust and Huber 2006; Lazareva et al. 2007). It is therefore possible that their depth perception in two-dimensional images is generally impaired unless motion-induced cues such as motion parallax are available. Unfortunately, there are no current studies exploring pigeons’ (or any other bird species) amodal completion with static three-dimensional objects; thus, the conclusions about the generality of monocular depth cues in avian vision must await further comparative research. Finally, there is no current evidence to suggest that insects or other invertebrates either recognize partially occluded objects or amodally complete them (Baum et al. 2014).
Human observers experience little difficulty in navigating the world filled with partially occluded objects, possibly due to their ability to spontaneously perceive partially occluded surfaces as complete. Although animals must undoubtedly face the same problems, the evidence for the ability of vertebrate animals to recognize partially occluded objects and to amodally complete them remains mixed; it is currently unknown whether invertebrates employ amodal completion or use occlusion as a depth cue.
- Fujita, K. (2006). Seeing what is not there: Illusion, completion, and spatiotemporal boundary formation in comparative perspective. In E. A. Wasserman & T. R. Zentall (Eds.), Comparative Cognition: Experimental explorations of animal intelligence (pp. 29–52). New York: Oxford University Press.Google Scholar