Visual Cues and Constancies

Visual cues and perception

When we look around, the information that we see is sent to the brain. Interestingly, the retina's input is two-dimensional, yet we can see the world in three dimensions. So, how does our brain know about depth?

Visual cues and perception: Howe we see the world

Depth perception is our ability to see the world in 3D. It allows us not only to see how long and wide objects are but also how far away they are. Understanding the distance of different objects from us is crucial for interaction with the environment.

There are two different types of depth cues: monocular depth cues and binocular depth cues. Cues that can be processed with one eye are called monocular cues. So, if you wore an eye patch, you would still be able to create a 3D representation of the environment. In comparison, binocular cues are when both eyes receive visual information to allow depth perception.

Visual constancies meaning

We can perceive certain features of physical objects as the same despite changes in how they appear, which are called visual constancies. Let's now unpack this definition.

The way that objects are presented on our retina changes constantly, yet we don't see. For example, the furniture in our room continually changes shape as we move and our perspective changes. Similarly, when the lighting changes, we can still perceive objects as having the same colour. We don't see changes in oranges or apple's colour when the sun goes down. This allows us to see a stable and constant image of the world.

We distinguish three types of visual constancies; shape, colour and size constancy.

Perhaps the most relevant visual constancies for depth perception are shape and size constancies.

Our brain correctly interprets the distortions of the rectangle's shape as being due to changes in our or the rectangles positioning and not as a change in the object's shape.

Size constancy is also very telling in terms of distance and depth. We have a sense of how tall or wide different objects are. Therefore we can use information about changes in their size as hints about their distance from us. For example, we generally know how tall a human is. Therefore if we see a human in our visual field looking much smaller, we can assume they are standing far away from us.

Visual cues types and examples

Let's now take a look at the different types of visual cues and how we experience them in real life.

Pictorial depth cues

Pictorial depth cues are all considered monocular. They can be depicted on 2D images and don't require any movement.

The first cue we'll describe is the height in plane. Elements of the visual scene that are the most elevated are interpreted as being the furthest from us. In contrast, objects that are lower down are interpreted as being closer to us. Consider the image below, the tree that appears the furthest away is also the most elevated.

Relative size is another important pictorial depth cue. The apparent size of an object relates to how far away it is from us. We know how big a tree typically is by comparing what we know about the tree size with the size that the tree appears to be. From this, we can estimate how far it is from us; this cue is also apparent in the image below. The tree that is the biggest appears to be the closest, while the tree that is the smallest appears the furthest away.

Trees standing alongside a path, Alicja Blaszkiewicz, StudySmarter Originals

Now let's look at the photo of the sunflower field. One cue we can use to assess the depth in this picture is occlusion. The sunflower that is the closest partially overlaps the one behind it. Typically, the view of objects that are further away will be mostly or entirely obstructed by nearer objects.

Another aspect of the picture that varies depending on distance is the texture gradient. We can see the fine texture in the centres of the sunflowers closest to us; this texture becomes smaller and less visible when looking at the sunflowers furthest away.

Sunflower field, freepik.com

Let's move on to the highway photo, where we can clearly see linear perspective. Since we are used to seeing how parallel lines look from a distance, we can quickly interpret the lines converging on this picture as a sign of distance rather than the road becoming more narrow.

Linear perspective on the example of a highway, freepik.com

Binocular depth cues

Binocular depth cues use the compares the information received from each eye. The differing images allow us to extract more information about the object's distance. Even though our eyes are close together, each looks at the world from a slightly different perspective.

This retinal disparity, so the difference in information to the retina in each eye helps us perceive the relative distance of objects. This results in stereopsis, which is the perception of depth due to binocular vision.

Examples of visual cues and constancies in visual illusions

All these cues make a lot of sense and seem to be accurate. So, why do they sometimes mislead us? Every so often, the cues can be ambiguous or in conflict with each other, leading to visual illusions. While pictorial cues are mostly accurate in real-life settings, artificially created images can use them to trick our perception.

Let's consider the Kanizsa Triangle illusion. The placement of the figures in this image makes us see a white triangle in the middle, which is not actually there because we fill the contours with another figure.

Kanizsa triangle, Fibonacci, CC BY-SA 3.0 , via Wikimedia Commons

On the other hand, the Ponzo illusion is caused by our reliance on pictorial depth cues. The two yellow lines that you can see on the image below are actually of the same length, even though the line, which is higher up in the picture, appears much larger.

The illusion is caused because we have different pictorial cues like apparent size, height in plane, texture gradient and linear perspective, all of which suggest that the yellow line higher up is relatively much longer compared to the other one.

• One of the lines is much shorter than the black lines parallel to it, and the other is longer; this gives us information about their apparent size.

• Another cue is the height in plane; the difference between the height of the two lines suggests that one is further away.

• The texture gradient created by the shorter black lines becomes smaller higher up on the picture, signalling information about distance. This also suggests that one of the yellow lines is much further away than the other.

• We can also see linear perspective created by the convergence of the two lines perpendicular to the yellow lines, indicating depth.

If this were a scene we saw in real life, we would be right, and the lines wouldn't be the same size!