This preview shows page 1. Sign up to view the full content.
Unformatted text preview: Chapter 5: Perception 1 Sensation and Empiricism
1. In sensation we studied stimulus features of the object and how they are registered in the brain as lines, edges, colors etc. 2. This idea of breakdown of stimulus features and then putting them back together through associations was put forth by empiricist school of thought. To them sensation was a passive process in nature. 3. However, as we discovered in the last chapter, sensation is an active process that manipulates information. 2 Perception
1. Process of sensation first manipulates information by breaking it down into features. 2. At a next stage, processing of information becomes more complex. Simple features combine together to form precepts that make objects, contexts, and events. How this achieved is not completely understood. 3. In this next level of processing one integrates, organizes, and interprets information, and is called perception.
3 Perception: Problems
1. It seem so natural and easy to open our eyes and start seeing and recognizing objects around us. But perception is indeed a complicated process. 2. First how do we interpret the object? How do we know that the apple is sweet, that it is edible, and that we can make jam out of it. 3. Even more difficult problem is how do we come about to see an apple as an entity on its own? One who has never seen an apple before readily recognizes it as say, one object. 4 Perception: Problems
4. To simplify some of our aforesaid problems regarding this complicated stage of information processing, we can break it down into three important area (questions) and ask ourselves: a. Where is the object to be perceived? b. What is the object doing? c. What is object? (most difficult question) 5 Distance Perception: Where is it? 6 Distance Perception: Cues
Distance perception is afforded by cues the stimulus context provides us and indirectly let us know how far an object is from us. We use both binocular and monocular cues to perceive depth. www.slrphotographyguide.com 7 Binocular Cues
1. When a comparison is generated by two disparate retinal images (retinal disparity) depth perception takes place. The difference in retinal images is due to different distances between each retina and the objects. Convergence
2. When two eyes move inward (towards the nose) to see a near object they converge and let the brain know that the object is close by. The opposite happens when they diverge and view a far away object. cnx.org 8 9 Monocular Cues
1. Monocular Cues include, object interposition. Car looks farther away than the mailbox, because of occlusion. Even simple objects (gray square) occluding one another give sense of depth. 10 Linear Perspective
2. Another way we perceive distance is to use linear perspective. Objects become small as they recede into distance. Linear perspective is built on converging lines that vanish a point. i.pbase.com 11 Relative Size
3. All things being equal, similar objects are perceived in depth if one of them becomes smaller in size. The smaller sized (relative size) penny looks farther away. 12 Texture Gradients
4. Uniformly textured surfaces produce texture gradients that provide information about depth. Figure A shows upward tilt at the corner, and Figure B a sudden drop. 13 Motion Parallax
5. Objects closer to a fixation point move faster and in opposing direction to objects farther away from a fixation point, which move slower and in the same direction. This is called motion parallax and provides depth cues. 14 Light and Shadows
6. Nearby objects reflect more light than distant objects. Given two identical objects, the dimmer one seems further away. When light coming form above (sun) shines an object, we perceive it as bulging. 15 Optic Flow
7. Optic flow is depicted as motion vectors (patterns of stimulation) across our retina. Optic flow information changes as we move either closer or farther away from an object. Expansion Contraction 16 The Role of Redundancy
1. Both binocular and monocular cues overlap in terms of information they provide. But why this redundancy? 2. Some cues operate more precisely under certain conditions than others, e.g., binocular cues are superior at distances under 30 ft; monocular cues better for much longer distances. 3. All cues are needed to maintain our ability to perceive distance across many situations in a dynamic environment.
17 Motion Perception What is it Doing? 18 Perception of Motion
1. Perception of motion involves time‐based events in which an object moves from point A to point B taking some time. 2. If an object moves too slow it may be perceived as stationary and if too fast no motion may be perceived. So there is a range of time in which when objects displace and are perceived as moving. A B 19 Perception of Motion: Retina
1. Certain cells in retina are direction specific cells, and will respond when stimulated by moving objects in one direction (say left to right), such cells are called motion detectors. 2. But there are many other complex ways and many other visual areas, through which we perceive motion. 20 Apparent Motion
Retinal motion is just simply the beginning of how we perceive motion. Think when a light spot briefly stimulates the retina at one place and then another spot, the result is apparent motion. The image does not move on the retina, but is perceived as moving. How? 21 Apparent Motion
Apparent motion, called phi phenomenon by Gestalt psychologists, suggested that this movement was produced in the brain (mind). So cells in the brain were generating this movement. 22 Eye Movement
When an object moves, the retinal image moves and we perceive motion, however when the object is stationary and the retinal image moves, due to eye movement, why don’t we perceive motion? 23 Eye Movement: Nervous System
1. Neuroscientists believe that as the object moves on the retina, our eye or head movements let the nervous system know and cancel (compensate) the effect. So we perceive a stationary object. 2. To test this idea, could we generate an illusion of movement in a stationary object by making our eyes or head free. The phenomenon produces an illusion of movement called induced movement.
24 Induced Motion
Illusions of induced movements include: a. The moon sailing through a stationary sky b. The train next to us moves; but we perceive ourselves as in motion. www.phili‐photo.com http://psychlab1.hanover.edu/Classes/Sensation/induced/ http://www.uni‐bielefeld.de/(en)/psychologie/ae/Ae01/forschung/indumo.html 25 Induced Motion: Reasons
1. Ptolemy (83–168) discovered induced motion. 2. Most important cue for this phenomena is the still “frame” which keeps our eye movements stationary. Moving bars generate movement on the retina and since the eye and retinal movements do not cancel each other we see a stationary object, midst bars, move in the opposite direction. Claudius Ptolemaeus 26 “Ptolemy” Form Perception What is it? 27 Form Perception
1. So far we have considered where an object is, what is it doing , but perhaps the most important question is, what the object is to begin with. 2. When we say, what is the object? We are simply saying how do we recognize the object? Recognition of objects is based on sometime color, at other times size, but most importantly it is the form. How is form of the object recognized? 28 Feature Detectors
1. Empiricists and modern day neuroscientists suggest that object form is decomposed into features processed by many feature detector cells in the visual system. 2. All features of the object are then compared to checklist of features in the brain (mind) to recognize the object. For example, We can decompose the letter “A” into at least three features (/ ‐ \) or line segments, compare it to A in our mind and recognize it.
29 Feature Detectors
1. Behavioral studies indicate that features have a special status in perception of form. 2. Treisman and colleagues (1980, 1986, 1988) have shown that visual search for a target among distracters is swift if the target has one distinct feature. The target “pops out”. Anne Treisman30 webscript.princeton.edu Organization in Perception
1. There is more to perception than merely cataloging object features. 2. We can see the giraffes and recognize them though they are exposed in many different ways. Features for all these giraffes differ, so how do we recognize them? 3. Is there a procedure that extracts a standard view from features analyzed?
31 Organization in Perception
4. Features may not readily accessible of say a partially occluded object (blue square), and yet we recognize it. 5. In still other cases, we recognize the object (man), process some of the features and completely filter others that make the scene. 32 Organization in Perception
6. So this suggests that recognition of an object only partially depends on features. Their absence, or intentional removal have little effect of recognition. 7. How do we then recognize the object by its form? Many experts believe that recognition of objects are based on how we organize features in them. 33 Organization in Perception
8. Tell me what you see? For some the meaningfulness of the stimulus slowly comes. Although the features of the stimuli are present they do not mean anything until the form is recognized. 34 Organization in Perception
9. Here is another example, you can easily recognize the stimulus below, however the essential features are missing to form letters. Suggesting that there more to form recognition than merely aggregation of stimuli. 35 Organization in Perception
10. One final example shows features (illusory contours) of Kanizsa triangles are completely absent and yet we can see triangles (form). 36 Gestalt School
1. Almost a century ago, Gestalt psychologists proposed that features alone or their aggregates could not result in recognition of perceptual objects. 2. Form of an object depends on organization of the entire pattern. The word Gestalt means “entire pattern” or “form”. 37 Perceptual Parsing
1. What does it mean to find “organization” within a figure? Many factors are needed to achieve this organization, perceptual parsing is one of them. 2. The banana in the background is divided into two parts covered by another banana in the front. We recognize it is the same banana. The two apples side by side are perceived as different, though their colors are the same. The process that achieves all that is called perceptual parsing.
38 farm3.static.flickr.com Perceptual Parsing
1. Perceptual parsing provides our perceptual system with groups of elements that can be combined and separated. 2. It does that by using a number of principles outlined by Gestalt psychologists. 3. Similarity and proximity are two principles that tie elements in a scene together. Similarity Proximity 39 Perceptual Parsing
4. Also good continuation and closure work as other principles to afford form perception. Good Continuation Closure 40 Figure & Ground
1. One last thing that is important for object recognition is figure‐ground relationship. 2. The object (figure) has definite shape or contour and seems closer and more dominant than the background. Perceiver determines the figure‐ ground relationship (see reversible figure). Figure‐Ground Relationship Reversible Figure 41 Different Perspective on Perception 42 Perspectives
1. We have looked at many aspects of perception, but need to have unified understanding of how perception takes place. 2. Experts think about perception at three broad levels: a. Classical approach – understanding perception at a global level. b. Process‐model approach – filling in details of how perception proceeds. c. Neuroscience approach – understanding 43 perception at the brain level. The Classical Approach to Perception 44 Classical Approach
Look at the Necker cube, you can either see view B or view C. And they keep fluctuating, just like the reversible figure on the left; vase or faces. Reversible Figure Necker Cube 45 Context
1. The two above (ambiguous) examples suggest that there is a lot that needs to said about how we organize and interpret information that reaches our senses. 2. In fact in many cases context affects our perception. THE CHT
46 Perceiving Constancy
The perceiver plays another important role in interpreting various perceptual constancies. So as the distal object (door) changes its shape, proximal image compensates some how and keeps the shape of the door constant (rectangle). Shape Constancy 47 Perceiving Constancy
Similarly we have size, brightness, color, and position constancies. Color Constancy
48 Size Constancy Higher‐Order Invariants
Gibson suggested that the object signals itself constancy information. The pattern of stimulus (size, shape) contains higher‐order invariants; these form relationship between retinal image and the object. J. J. Gibson (1904–1979) 49 Unconscious Inference
1. How do we achieve constancy when higher‐ order invariant relationships are available? 2. Helmholtz proposed that retinal image size and distance are inversely related and served as a principle to calculate constancy. So if the object is close we see it as bigger, and if far away, smaller. This calculation was carried out unconsciously (unconscious inference). 4 mm X 3 m = 12 2 mm X 6 m = 12 50 Illusions
1. Many times we are fooled by our senses. What we see is not really a true depiction of reality. Illusions or false perceptions provide us with special circumstances to study perception. Ponzo Illusion 51 Illusions
2. The length of the thin table is equal to the width of the table that looks square. The two yellow logs on the tracks are equal. And the color and brightness of square A and B are the same. 52 Objectivity of Perception
1. Study of illusion suggests that the objects may be misperceived, and if the perceiver is interpreting and organizing stimulus information then how does she know what really is “out there”? 2. Fortunately, human perception suffers from illusions infrequently. And when it does, several cues assist the perceiver to correct the illusions or become aware of them so that necessary steps be taken to interact with them.
53 Objectivity of Perception
3. Why did evolution design us to have illusions some of the time? Evolution designed perception malleable, for it interprets ambiguous environment all the time. Occasional misperception thus may be harmless especially if other factors correct the problem. 4. Our interpretation of the stimulus environment makes our perception efficient, and most of the time very logical as in terms of solving a problem (Rock, 1983).
54 Read Aloud
Aoccdring to a rscheeaech at Cmabrigde Uinervtisy, it deosn’t mttaer in waht oredr the ltteers in a wrod are, the olny iprmoetnt tihng is taht the frist and the lsat ltteer be at the rghit pclae. The rset can be a tatol mses and you can sitll raed it wouthit porbelm. Tihs is bcuseae the huamn mnid deos not raed ervey lteter by istlef, but the wrod as a wlohe. Rayner and colleagues (2006) found that the speed of reading text, like above, drops down slightly.
55 The Process‐Model Approach to Perception 56 Process‐Model Approach
1. The classical approach to perception broadly outlines how the stimulus world is achieved. However it does not say anything about how the perceiver makes perception take place. 2. In other words, what kind of perceptual steps are needed to achieve perceptions of size, shape, form, or constancy or for that matter illusions. 57 Hierarchical Arrangement
1. Experts believe one way to achieve that is through hierarchy of feature detectors in our perceptual system. 2. At the lower level of this hierarchy singular features like lines are detected, at the next level features organize lines into angles, and then at still higher level into shapes (square). Such hierarchical models are called feature nets.
58 Hierarchical Arrangement
3. In early hierarchical models activation flowed in one direction. From lower level feature detectors to higher level feature detectors. 4. More recent models propose that activation flows from features detected from stimuli (bottom‐up or data‐driven processes) and from our expectations, beliefs and our memories (top‐down or knowledge driven processes). 59 Top‐down Processing
5. If we are looking at the word “CAT” in dim light and cannot recognize it, top‐down processes help us perceive the word. 6. If we know that the word depicts an animal, MAT is filtered immediately. System then needs to decide whether it is a CAT or a RAT. Meek bottom‐up process registers a leftward curve and augments our perception for the 60 word CAT and not RAT. Sensory Memory
1. How long does it take for the stimulus information to pass from one level of feature detectors to another level and finally be recognized? 2. Our visual system holds information (sensory memory) for a very brief period of time (250ms) at the lowest level of processing and then passes it on to higher levels. 61 Features, Geons, & Meaning
1. So far we have looked at features of objects that are two dimensional. But the real world is 3D. How do we arrive at perceiving objects that have volume. 2. Biederman says through volumetric primitives or geons (short for “geometric icons”). 62 Features, Geons, & Meaning
3. Geons can be put together to form many different 3D shapes. 4. Although geons can represent the geometry of the object at hand, they do not help in identifying their meaning. 5. To decipher the meaning of the object it takes further intricate processing for its recognition.
63 Features, Geons, & Meaning
6. Luria (1973) for instance talks about a case where the patient described the geometry of spectacles, but when asked what it was, said a bicycle. 64 The Neuroscience Approach to Perception 65 Visual Pathway
1. Visual pathways begin at the retina and contain photoreceptors, rods and cones, that process brightness and color information. 2. Connected to them are two types of retinal ganglion cells; parvo and magno cells. 3. Parvo cells are sensitive to color, form and pattern, and magno cells are sensitive to brightness, motion and depth. 66 Retina and LGN www.sharp‐sighted.org Parvo: Color, form, and pattern Magno: Brightness, motion and depth
67 Visual Pathway
4. As the visual pathways progress more towards the occipital lobe, the parvo and magno streams get more and more segregated. 68 “What” and “Where” Systems
5. And finally end up in two streams processing “what” (inferotemporal cortex) the object is “where” (posterior parietal cortex) it is located. 69 The Binding Problem
1. How do these parallel streams process bits of information and recombine them? 2. Neuroscientists call this the binding problem: How various inputs processed by separate systems bind together into the correct representation of a single object. 3. One way this may be achieved is through neural synchrony. Neurons processing motion, color, shape, fire in synchrony; around 40 Hz, called the gamma band oscillation.
70 Perceptual Selection: Attention 71 Selection
1. What starts this synchrony? One possible mechanism is attention. 2. Attention causes selective control of orientation and selective looking. Attention is easy when target has one distinct feature. But is difficult when target represents multiple features. 72 Illusory Conjunctions
3. Unless “binding” occurs of all features in question, “illusory conjunctions” result – a feature from one object is assigned to another object instead. Rather than naming the target as a red “O” the responder says green “O” or red “V”. 73 Priming
4. How do we direct or select objects to attend to? 5. The answer is priming. Priming is based on our expectations and directs our attention to events making it more efficient. 6. Priming is an internal selection process that moves our mind’s eyes, even though our physical eyes remain stationary. 74 Perception: Attention Absent
Attention orients, binds, and when primed through expectation makes the process efficient. About half of the participants were blind to the man with the umbrella or in the gorilla suit (Simon & Chabris, 1999). www.perceptionweb.com 75 Other Modalities 76 Other Modalities
1. Hearing has much in common with vision. Sounds must be parsed like visual aspects of the stimulus. 2. Context works similarly in hearing – “Kiss this guy” versus “Kiss the sky” have identical sounds, but different contexts. Attention is a crucial determinant of what is heard. 3. Other modalities less rich for human beings than are vision and hearing.
77 Some Final Thoughts 78 Perceiver and Knower
1. Perception is not knowledge. Perception frequently contradicts what we know is “possible” or not. 2. Knowledge and perception are mutually influenced, but are not the same thing. The perceiver and the knower are distinct in many instances. www.indianchild.com 79 Questions
1. Why is it helpful for humans to be sensitive to many different, and sometimes redundant, distance cues? a. Different distance cues are important in different situations. b. Distance cues are usually misleading, and we need back‐up systems to prevent visual illusions. c. No distance cue, by itself, can provide interpretable information about an object’s distance. d. Each distance cue is only accurate for objects within a specific distance range.
2. Direction‐specific motion‐detecting neurons in the visual cortex are sensitive to motion because they fire in response to: a. changes in light intensity of a stimulus. b. a stimulus moving across the neuron’s receptive field on the retina. c. voluntary movements of the eyes. d. monocular distance cues.
3. Helmholtz argued that we achieve perceptual constancy in the absence of higher‐order invariant cues because we make ____________ about visual stimuli. a. visual constraints. b. visual illusions. c. unconscious inferences. d. conscious calculations.
4. Which of the following phenomena is an example of a knowledge‐driven, top‐down perceptual process? a. sensory memory. b. feature detection. c. proximal stimulus. d. priming effects. 83 Questions
5. Visual processing in the brain is characterized by _______________ processing, in which multiple processes occur simultaneously. a. stepwise. b. linear. c. parallel. d. hierarchical. 84 ...
View Full Document
This note was uploaded on 03/09/2010 for the course PSY 1073 taught by Professor Ahmad during the Spring '08 term at Henderson.
- Spring '08