Vision Science - depth perception
Sources of depth information: ocular information (accommodation and convergence), stereoscopic information (binocular disparity and Da Vinci stereopsis), dynamic information (motion parallax and accretion/deletion of texture), and pictorial information (all the rest).
26 flashcards · Biology · by lakesare
Name 2 ocular phenomena that let us percept depth. Sources of depth information: ocular information - ________________ and _________________ stereoscopic information dynamic information pictorial information
accommodation - when the lens changes thickens to focus light on the macula. convergence - extent to which the two eyes are turned inward (toward each other) to fixate an object. Because convergence depends on the observer using both eyes, it is a binocular source of depth information, unlike accommodation. Like accommodation, however, convergence provides information about the absolute distance to the fixated object.
[ocular information] Explain accommodation.
When ciliary muscles contract the lens in order to focus on the object. Beyond 6-8 feet, however, accommodation provides little or no depth information. At this distance the muscles that control the shape of the lens are already in their most relaxed state, so the lens cannot get any thinner. Although accommodation is generally considered to be a weak source of depth information (Hochberg, 1971), experimental results indicate that people use it at close distances. Observers are notoriously poor…
[ocular information] Explain convergence?
2 eyes are converged directly at the point so that light coming from it falls on the centers of both foveae simultaneously. Depth information from eye convergence. The angle of convergence between the two eyes varies with the distance to the object they fixate: smaller angles for objects far away and larger angles for objects nearby.
Does accommodation provide information about absolute or relative depth?
Absolute For accommodation to provide accurate depth information, the object must be in proper focus on the retina. This implies that the visual system must somehow "know" when an object is in focus. The best indication of proper focus is the presence of sharp edges rather than blurry ones. Thus, image blur is the effective stimulus that drives accommodation. Information about the blurriness/sharpness of edges is carried in high spatial frequencies: sharp edges contain more energy in high spati…
The convergence can be used to deduce absolute distance towards the object. Is it actually used this way by humans? This equation shows that if the distance between the two eyes is known, the angle of eye convergence could be used by the human visual system to determine distance to the fixated point.
It is, but only up to a distance of a few meters. At close distances the convergence angle changes rapidly for points that are only slightly different in depth, but at distances beyond about 2-3 meters, convergence changes very little as it approaches an asymptote (or limiting value) of zero degrees when the eyes are directed straight ahead to converge on a point at infinite distance. Controlled experiments have shown that it is (e.g., Hofsten, 1976), but only up to a distance of a few meters.
Name 2 stereoscopic phenomena that let us percept depth. Sources of depth information: ocular information stereoscopic information - ________________ and ________________ dynamic information pictorial information
binocular disparity - Because the positions of the eyes differ by a few inches, the two retinal images of most objects in the overlapping portion are slightly different. Da Vinci stereopsis
[stereoscopic information] What's binocular disparity?
Difference in images seen by the left and right eyes, resulting from the eyes’ horizontal separation.
[stereoscopic information] Experience binocular disparity :-)
First, hold up your left index finger at full arm's length and your right index finger at half arm's length. Then close your right eye, and align your two fingers using your left eye so that they both coincide with some distant point in the environment. Now, keeping your fingers in the same place and continuing to focus on the distant object, close your left eye and open your right. What happens to the images of your two fingers? They are no longer aligned either with the distant fixated point…
[stereoscopic information] What's crossed VS uncrossed binocular disparity?
Eyes are fixated on P by definition. C: crossed. F: uncrossed. When a point P is fixated, closer points (such as C) are displaced outwardly in crossed disparity, whereas farther points (such as F) are displaced inwardly in uncrossed disparity. By definition, point P falls on the foveae of both eyes and thus stimulates corresponding points. Now consider the projections of a closer point C while the eyes are still fixated on point P. They do not fall on corresponding retinal points,since the one…
[stereoscopic information] What point does not produce disparate images on the left and right retinae?
The horopter. The most obvious example is the fixation point, which falls on the center of both foveae. The set of all environmental points that stimulate corresponding points on the two retinae is called the horopter.
[stereoscopic information] Describe theoretical VS actual horopter. There are two ways of defining the horopter: theoretically by geometrical means and empirically by experimental means.
The theoretical horopter can be defined geometrically by projecting pairs of corresponding retinal points outward through the nodal point of the eye. Some of these pairs of lines intersect at a point in the environment,which, by definition, projects to corresponding retinal positions.The set of all such points constitutes the horopter. In the horizontal plane of the eye, the theoretical horopter is a circle,called the Vieth-Muller circle, which passes through the fixation point and the nodal po…
Why do we not experience double images (diplopia)? After all, every point in the external world that falls within the binocular portion of the visual field produces two retinal images, yet we seldom experience double vision.
Points near the horopter are fused perceptually into a single experienced image. The region around the horopter within which disparate images are perceptually fused is called Panum's fusional area. You can experience double images however if the amount of disparity is great enough, as when you cross your eyes by focusing on your nose. Panum's fusional area. Environmental points within Panum's area are fused perceptually into a single image.Points that are closer or farther produce double images…
[stereoscopic information] What are the limitations of usefulness of binocular disparity?
Effectiveness of binocular disparity is limited to a range of less than about 30 meters. Beyond this range, disparities become too small to detect reliably. Stereoscopic vision is also necessarily limited to the central region of the visual field where the two retinal images overlap. Not everyone who can see with both eyes has stereoscopic depth perception. Stereoblindness, as it is called, is relatively common,affecting from 5% to 10% of the general population (Richards,1970) and is usually ca…
There are two ways to do this: the crossed convergence method and the uncrossed convergence method.
Viewing the musing the uncrossed convergence method will reverse perceived depth, so that Figure 5.3.6A will look like Figure 5.3.6D, and Figure 5.3.6C will look like Figure 5.3.6B. Figure 5.3.6Binocular disparity and stereograms. If the twoimages in part A are stereoscopically fused with crossed conver-gence (see text for instructions), the circle and square will appearabove the page as indicated in part B. If the two images in part Care cross-fused, the circle and square will appear behind th…
[stereoscopic information] How is the correspondence problem solved during the stereopsis - shape-first, or correspondence-first? Flowcharts for two theories of stereopsis. (A) Theshape-first theory assumes that images in the two eyes arematched by comparing the results of two separate shape analyses.(B) The stereopsis-first theory assumes that stereopsis occurs be-fore shape analysis. Perception of random dot stereograms sup-ports the stereopsis-first theory (part B)
Correspondence-first :-) This is proved by random dot stereograms. Shape was assumed to be analyzed first, separately for the left and right images, so that theresults could be used to solve the correspondence problem. The rationale was that although it would be difficult to determine which point of light among thousands in one retinal image wentwith which point among thousands in the other, it would be easyto determine that, say, the tip of a German shepherd's nose in oneretinal image went wit…
How to construct the random-dot stereogram?
The left and right images are identical, except for a region that has been laterally displaced, and the unique regions revealed by the displacement.
Why does the square move with you in the random-dot stereogram? Phenomenal movement in random-dot stereograms. When the observer moves while viewing a random dot stereogram, the closer surface moves in the same direction (partB) because no texture elements are occluded or disoccluded during the motion, as would occur in real depth if both surfaces were stationary (part A).
As you move rightward, additional random dots would appear along the right edge of the square as they are disoccluded by the square, and other dots would disappear along the left edge as they are occluded by the square.But in the stereogram, the square is actually coplanar with the surrounding field, so neither occlusion nor disocclusion oftexture occurs when you move your head. Similarly, if the square were actually in front of the background, moving your head would generate motion parallax. N…
[stereoscopic information] What is Da Vinci stereopsis?
Da Vinci stereopsis. Depth information also arises from the fact that certain parts of one retinal image have no corresponding parts in the other image. Notice that there is a portion ofthe farther surface to the right of the closer surface that is seen bythe right eye but not the left. Similarly, there is a portion to theleft of the closer surface that is seen by the left eye but not theright.
Name 2 ocular phenomena that let us percept depth. Sources of depth information: ocular information stereoscopic information dynamic information - ________________ and _________________ pictorial information
motion parallax accretion/deletion of texture
A good summary of the similarity between binocular disparity VS motion parallax?
Binocular disparity - difference between 2 simultaneous retinal images, motion parallax - difference between 2 sequential retinal images. If you had moved your head laterally in the motion parallax demonstration by exactly the distance between your eyes, then the difference between the first and last retinal images would have been the same as the difference between your left and right images in the corresponding binocular disparity demonstration.
What distances is motion parallax VS binocular disparity effective on?
Motion parallax is also like binocular disparity in that it provides only relative information about depth. That is, it does not specify the actual distance to an object, but only how much closer or farther it is than the fixated one. Unlike binocular disparity, however, motion parallax can provide effective depth information at great distances. If you were to drive past two hills, for example, one of which was 2 miles away and the other 4 miles away, you would be able to determine which one wa…
[dynamic information] What's accretion/deletion of texture?
One further source of depth information arises from image motion events: the appearance (accretion) and disappearance (deletion) of texture behind a moving edge (Gibson, Kaplan, Reynolds, & Wheeler, 1969). Accretion and deletion of texture due to a moving observer are related to da Vinci stereopsis in the same way that motion parallax is related to binocular disparity. That is, accretion and deletion of texture are a dynamic source of depth information revealed over time that is conceptually th…
[pictorial information] Explain pictorial information.
The perspective projection that a perfect photograph achieves actually incorporates all of the pictorial depth cues we will discuss below because it captures all the information available in light falling on a single stationary retina
What's Alberti's window?
The basic idea is quite simple, as illustrated in Figure 5.5.2. To paint a 3-D scene accurately on a 2-D canvas, one merely needs to recreate the image that would pass through a flat window in place of the canvas. Taking this idea quite literally, one could recreate a realistic line drawing of a scene by viewing it through a pane of glass with one eye in a fixed position and tracing all the contours directly onto the glass with a grease pencil. If the glass were then placed over a uniform white…
Relationship between Alberti's window and pinhole camera?
The most interesting implication of the facts about perspective projection is that one can create convincing illusions of depth by employing Alberti's window. It is quite similar, in fact, to what a pinhole camera does when it captures an image on a flat sheet of film, as illustrated in Figure 5.5.3. The scene in Alberti's window is just a right-side-up version of the upside-down image that would be formed in a pinhole camera whose image plane is as far from the pinhole as the observer's eye is…
Name some pictorial informations that let us percept depth. Sources of depth information: ocular information stereoscopic information dynamic information pictorial information
Perspective Projection Convergence of Parallel Lines Position Relative to the Horizon of a Surface Relative Size Familiar Size Texture Gradients The 2nd motorcyclist is higher in the visual field