BCS 151 - Exam2 - Lecture notes 7-13 PDF

Preview

Summary

Information on Professor Tadin's second exam...

Description

Montalbano BCS 151 BCS 151 – Perception & Action Exam Two Study Guide Vision Grouping V1 responses to natural images  Early visual analysis is local

 

Visual input is analyzed at different spatial scales (i.e. spatial frequency bands) Different orientations and spatial scales, receptive fields take the picture apart and look at different features

The binding problem  We see objects not spatial frequencies, but how are local, spatially distributed image descriptions linked to achieve meaningful representations of objects?  Camouflage – makes the binding problem harder for other animals  Gestalt psychology – the whole is more than the sum of its parts o Vision exploits regularities in the environment and knowledge from experience to “solve” the binding problem  Unified = grouping  Segregated = figure/ground segmentation Perceptual grouping  Gestalt principles of grouping o Similarity  Lightness, color, orientation, texture o Proximity  The distance between objects affects how you see them (ex. rows vs. columns) o Common region  Physically grouping objects into regions overrides proximity o Element connection  Connecting objects with lines overrides proximity o Good continuation

Montalbano BCS 151 

o o

o o

The brain perceives a full line as long as the angles match even if the line is covered by an object Symmetry  Objects are grouped by being symmetrical Closure  “closing” seemingly random objects will allow one to perceive a more normalized picture Simplicity  We see the world in terms of simple objects rather than irregular ones Common fate  Things that move together group together

Perceptual segregation  Figure-ground segregation – determining what part of environment is the figure so that it “stands out” from the background  Border ownership – border belongs to the object in front, the edge is perceived as belonging to the person  Lower objects tend to be seen as the figure  Figure/ground ambiguity – the ability to perceive two different sets of figure and ground in the same picture Neural mechanisms of grouping and segregation  The brain is biased to find continuous shapes because certain neurons have receptive fields that respond more strongly to continuous shapes  Contextual modulation – stimuli outside of a neuron’s receptive field can affect neural firing  If the preferred feature is part of a larger shape, neuron firing increases  Picture  o On the top, the neuron is responding to the figure o On the bottom, the pattern is not part of the figure. So, the neuron may even receive inhibitory signals since it’s responding to something not as important o Neurons responding to objects fire more Color Why color?  Color is pleasing  Color affects mood  Color is useful – makes things conspicuous or inconspicuous  Color signals edibility/ripeness of food

Montalbano BCS 151 What is color?  White light of the sun can be decomposed into 7 basic colors  Colored light could be recombined to create white light  Basic color – one that cannot be further decomposed upon passing through a prism  Basic colors can be recombined to create new colors  Metamers – a perceived matching of colors with different spectral power distributions  The wavelengths reflected from an object characterize the visual stimulus and are (in part) responsible for our subjective sensation of color  A wavelength cannot be reflected if it isn’t in the ambient light  Additive is for mixing lights  Subtractive is for mixing pigments (i.e. paint) Trichromatic Theory  Each cone contains one of three different photopigments  The ensemble of cone signals carry information about wavelength of light  Univariance principal – a response of a single cone type cannot tell you about the color of the stimulus; it confounds intensity and wavelength  Trichromatic theory – o Researchers found that by mixing only three primary lights they could create the perceptual experience of all visible colors o We have three different types of photoreceptors, each most sensitive to a different range of wavelengths  Retinal color blindness – o Dichromacy: missing L or M type pigment (red-green deficiency)  L-type = Protanopia, M-type = Deuteranopia o Dichromacy: missing S-type pigment = Tritanopia Opponent-process Theory  Some aspects of our color perception are difficult to explain by the trichromatic theory alone o Ex. Afterimages – if we view colored stimuli for an extended period of time, we will see an afterimage in a complementary color  To account for complementary after images, Herring proposed that we have two types of color opponent cells: red-green and blue-yellow  In the 1950s and 60s opponent neurons were discovered – neurons in the retina and LGN that respond with an excitatory response to one color and with the inhibitory response to light from another color  Two theories together: o Trichromatic – very beginning of visual system, in the receptors of the retina, respond with different patterns to different wavelengths o Opponent-process – later in the visual system, from ganglion cells on, neurons integrate the inhibitory and excitatory signals from the receptors Effects of Context of Color Perception  Perceived colors depend on neighboring colors  Vision depends on context Synesthesia  A condition in which one associates objects with a sensory perception Depth Non-Stereo

Montalbano BCS 151 Depth perception  Absolute distance – how far from you is an object in the environment  Relative distance – how far away an object is relative to other objects in the visual field  3D surface layout – recovering orientation at a distance  Object shape = slant, tilt, and curvature  The optical projections of objects are inherently ambiguous  Inverse problem (the depth ambiguity) – same retinal image can correspond to an infinite number of real-world objects  An impossible figure – this illusion is created by having accidental co-termination of the object’s edges  Ames room – the trapezoidal shape of the room causes an accidental parallel alignment of the back wall when viewed through a small hole  Accommodation (a monocular depth cue) – is derived by image blur so that the output of high spatial frequency channels is maximized Monocular Depth Cues  Occlusion – a condition when a nearer object visually occludes at least a portion of a more distant object  Size – retinal image of an object depends on its size and distance; if you know the size, you can estimate the distance  Perspective o Geometric perspective – (linear perspective) convergence of lines that results in perceived depth in a 2D scene  Apparent distance can have a strong effect on apparent size o Texture perspective – the density of the surface/object texture increases with distance, providing a depth cue  Visual cliff – not crossing the “cliff” is understood to indicate the ability to perceive depth o Aerial perspective – objects in the distance appear less clear, more blurred and low contrast because of atmosphere and pollution  Shading – a gradient in the reflected light on a 3D object that gives cues about object’s 3D shape o Convex vs. concave shapes (convex is brighter on top, concave is brighter on the bottom) o Information from shadows or indirect illumination can override the bias to perceive objects to be in contact with the ground  Motion parallax – differences in relative motion of objects located at different distances from the observer Depth Stereo Stereoscopic Vision  Stereopsis – 3D vision resulting from slight differences in left and right eye images, arising because the two eyes view the world from slightly different perspectives Perception of Stereo Depth  Requires two 2D images taken from slightly different positions, and presented one to each eye

Montalbano BCS 151 

Predators typically have both eyes on the front of their heads and consequently have large binocular visual fields (have better depth perception)  Prey typically have one eye on either side of their heads and consequently have small binocular visual fields (can see almost everything around them)  Stereoscopes  Anaglyphs – the red and blue lenses filter the two projected images allowing only one image to enter each eye  Polarized glasses – only pass light whose oscillations are oriented in a particular direction  Active shutter glasses  Eye crossing  Autostereograms Retinal Disparity  A slight separation between retinal images of objects in left and right eyes  Retinal disparity depends on the object’s distance relative to the fixation point, and as such, it provides the geometric basis for stereopsis  Retinal disparity’s magnitude increases with depth differences  Horopter – the line connecting points (objects) that produce corresponding retinal points in the two eyes  Retinal coordinates – the position of the object on each eye  Disparity – slight differences in positions of “features” in left and right eye views  Corresponding retinal points = no disparity  Objects that are not on the horopter will be projected to non-corresponding points on the two retinas  Non-corresponding retinal points o In front of the horopter = crossed disparity o Behind the horopter = uncrossed disparity  The magnitude of disparity is defined as the difference in retinal coordinates between the optical projections of that point in the left and right eye  Disparity information provides information about relative depth Correspondence Problem  The brain has to match images from two eyes to compute depth  How does the brain “solve” this problem? o Derive the solution that maximizes the overall number of matches – i.e. that is most globally consistent Chromostereopsis  Short wavelengths refract more than long wavelengths  Thus, when light is refracted through the lens of the eye, blue and red light will be focused at slightly different positions on the retina  This creates a disparity between each eye’s view that is perceptually interpreted as a difference in depth Binocular Rivalry  When images in two eyes are radically different and a combined stereo cannot be found Motion General Necessary Motion  The retinal image is constantly moving or jittering

Montalbano BCS 151 o Saccades and microsaccades  If you paralyze your eyes, your vision will disappear Figure-Ground Segregation  Motion breaks even perfect camouflage  Motion makes depth appear even where there is none  Can make a 2D image 3D by playing two images in slightly different positions over in rapid succession Time to Collision  Optic flow – the motion of all surface elements from the visual world o Tells your brain how fast you are moving o Ex. Feels like your train is moving due to optic flow when really your train is still, and it is the train next to yours that is moving  Motion parallax – the objects closer to the train look like they’re moving backwards, the objects far away will look like they are moving in the same direction as you  Optic flow can be used to negotiate curves while steering a car Perceiving Others  Lights on joints example o Can gain a sense of emotion from movement alone o If specific injury to a certain part of the brain, there will be no structure perceived and it will just look like a bunch of dots moving Balance and Posture  The optic flow pattern is used to maintain balance while standing upright  Ex. Child in a moving room will counteract the apparent motion and fall over Perception of Moving Objects  What is motion? o Aristotle – motion is an independent property o Democritus – motion is the presence of objects in certain locations at certain times o In perception – there are several conditions that we perceive motion without change in location  Apparent Motion – producing motion perception from successively exposed static images o Ex. Flashes of dots appear to move left to right o Ex. Wagon wheel illusion  Aperture Problem – without the aperture the object clearly moves diagonally, but when looking through an aperture it appears to move up and down o The motion we perceive in the aperture is perpendicular to the actual motion (?)  Space time plots o Can depict motion as orientation, the brain can detect orientation really well  Eye movements – since our eyes and bodies are in constant motion the presence of motion on the retina does not require that motion actually occurred in real life

Montalbano BCS 151 

Corollary discharge – a copy of a motor command that is sent to the muscles to produce a movement o If the retinal motion is as expected due to eye movements, then no motion is perceived

Motion Neural Mechanisms Motion Pathway  Dorsal stream, action pathway, M pathway, where pathway  Retina to LGN – M cells respond strongly to rapid changes in intensity, ideal for conveying motion information to the brain  V1 – motion is detected, complex cells  MT – the motion area where detected motions are interpreted  MST – complex motion processing Motion Detection (V1)  Motion is detected by directionally selective complex cells  Motion detector (Reichardt detector) o The delay is the key part so that the inputs arrive at the same time o Directionally sensitive  Neurons can be directionally tuned Area MT  The interpretation of motion  The response of one neuron is ambiguous. Combination of outputs among many neurons is necessary to resolve ambiguity o Integration occurs in area MT which has larger receptive fields than V1  Ambiguities – the aperture problem, component vs. pattern motion  MT is located in the back of the brain in the occipital cortex  Akinetopsia – the inability to perceive motion  Without area MT we don’t see motion  Stimulation of MT results in the perception of motion o Ex. Monkeys signaled for downward motion when that was stimulated even when perceiving upward motion Area MST  Complex motion processing  Optic flow – expansion vs. contraction  Superior Temporal Sulcus (STS)  Motion aftereffect – waterfall example...