6. Space perception and binocular
vision
Here: ability to perceive & interact w/ the structure of space.
Geometry of real world = Euclidean.
Euclidean = parallel lines remain parallel, objects stay the same size, angles of a triangle add to 180
degrees, etc.
Geometry of retinal images, though, is not Euclidean. The geometry becomes non-Euclidean when 3D
images are projected onto curved, 2D surface of the retina. Parallel lines don’t always stay parallel, sizes
changes & triangles as well need to reconstruct the Euclidean world from a non-Euclidean input.
Our visual xpce = reconstruction of the world based on 2 non-Euclidean inputs (two retinal images).
They always differ.
Binocular visual fields better chance for predator animals (ex: human) to spot small, fast-moving
objects in front of them (dinner). Prey animals wide visual fields (monitor the whole scene for
predators).
Binocular summation = combination of signals from each eye in ways that makes performance on many
tasks better w/ both eyes than w/ either eye alone.
Binocular disparity = differences between the 2 retinal images of the same scene. Disparity = basis for
stereopsis, a vivid perception of the 3-dimensionality of the world that is not available w/ monocular
vision.
Stereopsis = ability to use binocular disparity as a cue to depth.
Monocular depth cue = a depth cue available even when the world is viewed w/ one eye alone.
Binocular depth cue = depth cue that relies on info from both eyes. Stereopsis = primary example.
Convergence & ability of 2 eyes to see more of an object than one eye sees = binocular depth cues.
Egocentric space = everything is represented depending on yourself (your perception, etc).
Monocular cues to three-dimensional space
We use depth cues to infer aspects of the 3D world from our 2D retinal images (not perfect
reconstruction). What cues does the vs use to infer depth relations, and how do we use those cues to
create a representation of the 3D world?
Occlusion
Regarding object recognition: occlusion = cue to the presence of an otherwise invisible edge. If an
edge suddenly stops, why does it stop? Guess something else gets in the way, hiding
it from our view.
Occlusion = depth cue. Gives info about relative position of objects. Wrong only in the case of accidental
viewpoint (triangle, square & circle = not in front of each other but have weird shapes). We also rather
see it that way because triangle, circle and square are simpler shapes than the one they are if they are
not actually in front of each other.
Occlusion = cue to relative depth order in which, for ex, one object obstructs the view of part of another
object.
Occlusion = nonmetrical depth cue. Doesn’t allow to assess size of objects (is the triangle smaller than
the square or larger but further away?). A Metrical depth cue, on the other hand, would provide info
about distance in the 3rd dimension.
Nonmetrical depth cue = depth cue that provides info about depth order (relative depth) but not depth
magnitude (his nose is in front his face).
vision
Here: ability to perceive & interact w/ the structure of space.
Geometry of real world = Euclidean.
Euclidean = parallel lines remain parallel, objects stay the same size, angles of a triangle add to 180
degrees, etc.
Geometry of retinal images, though, is not Euclidean. The geometry becomes non-Euclidean when 3D
images are projected onto curved, 2D surface of the retina. Parallel lines don’t always stay parallel, sizes
changes & triangles as well need to reconstruct the Euclidean world from a non-Euclidean input.
Our visual xpce = reconstruction of the world based on 2 non-Euclidean inputs (two retinal images).
They always differ.
Binocular visual fields better chance for predator animals (ex: human) to spot small, fast-moving
objects in front of them (dinner). Prey animals wide visual fields (monitor the whole scene for
predators).
Binocular summation = combination of signals from each eye in ways that makes performance on many
tasks better w/ both eyes than w/ either eye alone.
Binocular disparity = differences between the 2 retinal images of the same scene. Disparity = basis for
stereopsis, a vivid perception of the 3-dimensionality of the world that is not available w/ monocular
vision.
Stereopsis = ability to use binocular disparity as a cue to depth.
Monocular depth cue = a depth cue available even when the world is viewed w/ one eye alone.
Binocular depth cue = depth cue that relies on info from both eyes. Stereopsis = primary example.
Convergence & ability of 2 eyes to see more of an object than one eye sees = binocular depth cues.
Egocentric space = everything is represented depending on yourself (your perception, etc).
Monocular cues to three-dimensional space
We use depth cues to infer aspects of the 3D world from our 2D retinal images (not perfect
reconstruction). What cues does the vs use to infer depth relations, and how do we use those cues to
create a representation of the 3D world?
Occlusion
Regarding object recognition: occlusion = cue to the presence of an otherwise invisible edge. If an
edge suddenly stops, why does it stop? Guess something else gets in the way, hiding
it from our view.
Occlusion = depth cue. Gives info about relative position of objects. Wrong only in the case of accidental
viewpoint (triangle, square & circle = not in front of each other but have weird shapes). We also rather
see it that way because triangle, circle and square are simpler shapes than the one they are if they are
not actually in front of each other.
Occlusion = cue to relative depth order in which, for ex, one object obstructs the view of part of another
object.
Occlusion = nonmetrical depth cue. Doesn’t allow to assess size of objects (is the triangle smaller than
the square or larger but further away?). A Metrical depth cue, on the other hand, would provide info
about distance in the 3rd dimension.
Nonmetrical depth cue = depth cue that provides info about depth order (relative depth) but not depth
magnitude (his nose is in front his face).
