Depth cues are the visual, optical, geometrical and perceptual information used to interpret the distance, position, size, shape and spatial relationships of objects within three-dimensional space.
Human vision does not directly measure the absolute distance of every point in a scene. Light arriving at a particular position on the retina specifies a direction, but an object point located at many possible distances along that direction could potentially produce a corresponding retinal stimulation. The visual system must therefore use additional information to estimate depth and distance.
These sources of spatial information are commonly called depth cues. They include monocular cues available to one eye, binocular or stereoscopic cues produced through the use of two eyes, oculomotor information, movement, perspective phenomena and environmental optical effects.
Depth perception is therefore not produced by one single mechanism. Different cues may operate independently or together, and a convincing impression of three-dimensional space can be produced even when many of the available cues are absent.
What Are Depth Cues?
A depth cue is information within a view, image or visual process that contributes to the estimation or interpretation of spatial depth.
Depth cues can provide information about:
- relative distance;
- apparent size;
- position;
- orientation;
- shape;
- surface structure;
- overlap and spatial order;
- movement through space; and
- the three-dimensional organisation of a scene.
Some cues arise primarily from perspective geometry, others from the behaviour of light, others from movement, and others from the physiology and perceptual processing of the human visual system.
Why Are Depth Cues Necessary?
The retina receives a projected optical image of spatial reality. It can register the direction from which light arrives, but the retinal image does not directly provide a unique absolute distance for every object point.
This creates an important problem of correspondence between a two-dimensional retinal projection and the three-dimensional spatial reality from which it originated.
The visual system helps overcome this problem by combining many different forms of spatial information. Depth perception is consequently an interpretation of structured optical and visual information rather than a simple direct measurement of physical distance.
This is one reason perspective is fundamental to vision: many of the apparent changes that occur with viewpoint, distance and orientation provide information that helps us understand the spatial structure of the world.
Monocular and Binocular Depth Cues
Depth cues can be divided broadly into monocular and binocular or stereoscopic groups.
Monocular depth cues can operate when a scene is viewed with a single eye. They include many of the familiar perspective phenomena found in natural vision, photographs, paintings, drawings, cinema and computer images.
Binocular depth cues depend upon relationships between the two eyes, particularly differences between the images received from their slightly separated viewpoints.
The two groups should not be treated as competing explanations. Human spatial vision normally combines numerous cues according to the scene, viewing distance and task involved.
Monocular Depth Cues
Monocular depth cues are particularly important because they can provide a strong impression of three-dimensional space without stereoscopic vision.
Important monocular cues include:
- perspective of form;
- diminution of size;
- foreshortening;
- convergence of parallel lines;
- occlusion and overlapping contours;
- surface texture and texture gradients;
- aerial perspective;
- changes of colour and contrast;
- light and shade;
- height in the visual field;
- horizon relationships;
- known or interpreted object size;
- focus and acuity effects;
- motion parallax;
- optic flow; and
- changes of visible form with viewpoint.
Many of these cues are themselves perspective phenomena. This demonstrates the close relationship between depth perception and perspective.
Perspective as a Depth Cue
Perspective is one of the most important sources of monocular depth information.
As objects and spatial structures extend away from an observer, their projected appearances change systematically. Objects may diminish in apparent size, dimensions become foreshortened, parallel directions can converge, surfaces change shape with aspect and spatial intervals become progressively compressed.
The visual system can interpret these structured changes as evidence of spatial depth.
A familiar example is a pair of railway tracks. The rails remain parallel in physical space but appear progressively closer together with distance and converge towards a vanishing point. At the same time, sleepers, markings and other repeated features diminish and become more closely spaced in the image. Together these perspective phenomena provide powerful information about recession and spatial extent.
Diminution of Size
Diminution of size is a major perspective depth cue. Under appropriate viewing conditions, equal-sized objects located progressively farther from the observer produce progressively smaller projected images.
Where the actual or probable size of an object is known, apparent size can contribute strongly to judgements of relative distance.
Relative size does not require the presence of a visible vanishing point. Two recognisably similar objects can provide useful depth information simply through their different projected sizes.
Foreshortening and Aspect
Foreshortening provides information about the orientation and recession of objects and surfaces.
The Perspective Research Centre distinguishes two related forms: aspect foreshortening, resulting from viewing or projection angle, and perspectival foreshortening, associated with apparent size diminution through depth.
A surface viewed frontally presents a different apparent shape from the same surface viewed obliquely. These transformations help indicate its orientation relative to the observer and contribute to the perception of three-dimensional form.
Occlusion and Overlapping Contours
Occlusion, also called interposition or superposition in some contexts, is one of the simplest and strongest depth cues.
When one object partially blocks another from view, the occluding object is normally interpreted as being nearer to the observer while the obscured object is interpreted as lying farther away.
Overlapping contours therefore provide information about the relative spatial ordering of objects even when their absolute distances remain unknown.
Texture Gradients
Texture can provide powerful information about the orientation and recession of surfaces.
Repeated surface details tend to become progressively smaller and more densely packed in the projected image as a textured surface extends into depth. This structured change forms a texture gradient.
Road surfaces, fields, brickwork, tiled floors and repeated architectural details can therefore communicate depth without requiring stereoscopic information.
Aerial Perspective as a Depth Cue
Aerial perspective provides depth information through systematic optical changes associated with distance and atmospheric transmission.
Distant forms may become less distinct, show reduced contrast and undergo changes of apparent colour. The resulting gradients can help establish the relative depth of different parts of a scene.
Aerial perspective therefore differs from purely geometrical line perspective, although the two normally operate together in natural views.
Light and Shade
The distribution of light and shade provides information about surface orientation, curvature, relief and spatial position.
Gradients of illumination and shadow can help distinguish projecting from receding surfaces and contribute strongly to the appearance of three-dimensional form.
Chiaroscuro and other controlled lighting methods exploit these relationships in graphical representation, while comparable optical effects arise naturally in physical scenes.
Height in the Visual Field and the Horizon
The position of objects relative to the horizon and within the visual field can also contribute to depth interpretation.
For objects resting on a common ground plane, more distant positions are normally projected progressively nearer to the horizon line. This relationship provides information about spatial recession and ground-plane position.
The horizon is therefore not simply a drawing convention. Within appropriate geometrical configurations, it forms part of the structured perspective information from which spatial relationships can be interpreted.
Motion Parallax
Motion parallax is a depth cue produced by changes in relative viewing direction as either the observer or objects move.
Objects at different distances change their apparent positions and angular relationships at different rates as viewpoint changes. These differences provide information about relative depth.
Motion parallax is particularly important because it allows spatial information to be obtained through a sequence of changing monocular views rather than through two simultaneous binocular views.
Optic Flow and Movement through Space
As an observer moves through an environment, the pattern of visual information changes continuously. This changing structure is commonly described through optic flow.
Nearby and distant parts of a scene transform differently during movement, providing information about direction, relative distance and the organisation of surrounding space.
Movement therefore adds an important temporal dimension to perspective. Human spatial perception is not limited to interpreting one fixed retinal image but continually uses changing views as the observer, eyes, head and surrounding objects move.
Focus, Acuity and Depth
Changes in focus, clarity and visual acuity can also contribute to depth interpretation.
Objects and details at different distances may differ in their degree of sharpness according to the optical system, focus condition, depth of field and resolving capability of the eye or imaging instrument.
Loss of form, outline clarity or fine detail with distance can consequently act as additional spatial information, especially when combined with other cues.
Accommodation
Accommodation is a physiological depth cue associated with changes in the optical power of the eye as it focuses upon objects at different distances.
Accommodation belongs to a different class of information from pictorial cues such as relative size or convergence of represented lines. It arises from the functioning of the visual apparatus itself rather than solely from the visible geometry of a scene.
Binocular Disparity and Stereopsis
The two human eyes occupy slightly different positions and therefore receive slightly different views of spatial reality.
The differences between corresponding image features are known broadly as binocular disparity. The visual system can use these differences to obtain stereoscopic information about relative depth, a process commonly called stereopsis.
Stereopsis can provide particularly precise relative-depth information at close range. Its effectiveness diminishes as viewing distance increases because the angular differences between the two eye views become progressively smaller.
Binocular disparity is therefore an important depth cue, but it is not the sole basis of human three-dimensional vision.
Binocular Convergence
Convergence, or binocular vergence, provides another source of distance information.
When the eyes fixate objects at different distances, their relative angles of alignment change. The visual system can use information associated with this changing vergence as one contribution to distance estimation.
Convergence should be distinguished from stereopsis. Convergence concerns the relative orientation of the eyes, whereas stereopsis principally concerns differences between the two retinal images.
Do We Need Two Eyes to Perceive Depth?
No. Binocular vision provides important depth information, but stereoscopic vision is not required for a convincing perception of three-dimensional space.
Closing one eye does not make the spatial world suddenly appear flat. Monocular perspective, relative size, occlusion, motion, texture, shading and many other cues continue to provide extensive information about depth.
Binocular stereopsis is particularly useful for fine relative-depth judgements at close range, while larger and more distant environments rely increasingly upon monocular and perspective information.
This distinction is fundamental to understanding why a flat photograph, painting, drawing, film or television image can still present an extremely convincing impression of three-dimensional space.
Depth Cues in Pictures and Photographs
A conventional picture or photograph cannot reproduce every depth cue available when viewing physical reality directly.
Nevertheless, it can contain many powerful monocular cues, including diminution of size, foreshortening, convergence, occlusion, texture gradients, lighting, aerial perspective, horizon relationships and known object scale.
The visual system interprets these features in relation to its experience of spatial reality. A flat image can therefore evoke a compelling impression of depth even though the physical picture surface itself remains two-dimensional.
This is one of the fundamental principles underlying graphical perspective, photography, cinema and many forms of digital imaging.
Depth Cues and Linear Perspective
Linear perspective deliberately reproduces several geometrical depth cues on a flat image surface.
These can include:
- diminution of apparent size with distance;
- aspect changes;
- foreshortening;
- convergence of parallel directions;
- vanishing points;
- horizon relationships;
- progressive compression of spatial intervals; and
- perspective distortion of grids and repeated forms.
Linear perspective does not automatically reproduce every optical depth cue. Colour change, loss of contrast, atmospheric effects, fine-detail loss and complex lighting may require additional graphical or optical processes.
Depth Cues and 3-D Images
The expression 3-D image can describe several very different kinds of image and viewing system.
A monocular perspective picture can generate a strong appearance of three-dimensional space using perspective and other monocular cues. A stereoscopic system adds binocular disparity. Multi-view, holographic and virtual-reality systems may add further cues associated with changing viewpoint, vergence, parallax, focus and movement.
No existing representational method necessarily reproduces every depth cue available in natural physical reality. The degree of spatial realism therefore depends upon the complete system of image formation, presentation and observation rather than upon one cue alone.
Depth-Cue Conflicts and Perspective Illusions
Depth cues do not always agree. When different sources of spatial information indicate incompatible interpretations, depth-cue conflicts can occur.
Perspective illusions can deliberately exploit these conflicts. Converging lines, altered room geometry, unusual scale relationships, misleading shadows, ambiguous contours or conflicting object sizes can cause the visual system to interpret distance, position, shape or size incorrectly.
Such illusions demonstrate that depth perception is an active process of interpretation. The visual system combines available information and attempts to derive a plausible spatial organisation, but that interpretation can be manipulated or become ambiguous.
Depth Cues and Visual Perspective
Within Perspective Category Theory, human depth perception is principally associated with Visual Perspective Type 2: the human retinal and perceptual subclass of the wider category of Visual Perspective.
Visual Perspective Type 2 concerns the appearances produced when a person views physical or represented spatial reality from a particular viewpoint. These appearances are shaped by scene geometry, viewing distance, direction, optical projection, binocular vision, eye and head movement, acuity and other visual processes.
Depth cues are therefore not an isolated subject outside perspective. They are among the principal mechanisms through which perspective phenomena become useful information about three-dimensional spatial reality.
Depth Cues — Frequently Asked Questions
What are depth cues?
Depth cues are visual, optical, geometrical, physiological or perceptual information that helps the visual system interpret the distance, position, size, shape and spatial relationships of objects within three-dimensional space.
What are monocular depth cues?
Monocular depth cues are cues available when viewing with one eye. They include perspective, relative size, foreshortening, occlusion, texture gradients, aerial perspective, light and shade, horizon relationships, motion parallax and other changes in apparent form.
What are binocular depth cues?
Binocular depth cues arise from the use of two eyes. Important examples include binocular disparity or stereopsis and changes in binocular convergence or vergence.
Is perspective a depth cue?
Yes. Perspective provides several important monocular depth cues, including diminution of size, foreshortening, convergence, vanishing-point relationships and systematic changes in the apparent geometry of objects and surfaces with distance and viewing angle.
Can we perceive depth with one eye?
Yes. Monocular depth cues provide extensive information about three-dimensional space. Binocular vision improves particular kinds of relative-depth judgement, especially at closer distances, but it is not necessary for the general perception of spatial depth.
Why do flat pictures look three-dimensional?
Flat pictures can contain many of the same monocular depth cues present in natural vision, including relative size, perspective convergence, occlusion, foreshortening, texture, shading, aerial perspective and horizon relationships. The visual system interprets these cues as evidence of represented three-dimensional space.
Is stereoscopic vision necessary for 3-D perception?
No. Stereopsis is an important binocular depth cue, but convincing three-dimensional appearances can be produced entirely from monocular information. Paintings, photographs, conventional films and ordinary television images demonstrate this every day.
Depth Cues and the Understanding of Perspective
Depth perception demonstrates why perspective cannot be understood simply as a method for drawing converging lines. Human spatial vision depends upon a much larger system of geometrical, optical, environmental, physiological and perceptual information.
Diminution, foreshortening, occlusion, texture, aerial perspective, movement, binocular disparity, convergence, focus and many other cues combine to transform projected visual information into an experienced three-dimensional world.
The study of depth cues therefore provides an important bridge between perspective geometry, optics, human vision, visual perception and the representation of three-dimensional space.