Chapter 3: Visual Perception

Some Questions of Interest How can we perceive an object like a chair as having a stable form, given that the image of the chair on our retina changes as we look at it from different directions?

Some Questions of Interest What are two fundamental approaches to explaining perception? What happens when people with normal visual sensations cannot perceive visual stimuli?

Perception Is… The process of recognizing, organizing, and interpreting information How do you recognize these items?

Basic Concepts Distal object Informational medium Proximal stimulation Grandma’s face Informational medium Reflected light from Grandma’s face Proximal stimulation Photon absorption in the rod and cone cells of the retina Perceptual object

Perceptual Basics Sensory adaptation Our senses respond to change Occurs when sensory receptors change their sensitivity to the stimulus Constant stimulation leads to lower sensitivity Our senses respond to change Ganzfeld effect

Perceptual Illusions Sometimes we cannot perceive what does exist Sometimes we perceive things that do not exist INSERT FIG 3.3 & FIG 3.4

Perceptual Illusions Sometimes we perceive what cannot be there INSERT FIG 3.5

Perceptual Illusions Some other illusions to explore: http://www.michaelbach.de/ot/#history

Our Visual System Light travels through the eye and focuses on the retina Electromagnetic light energy is converted into neural electrochemical impulses INSERT FIG 3.7

Our Visual System Three main layers of neural tissue in retina Ganglion cells Amacrine cells, horizontal cells, bipolar cells Photoreceptors Rods and cones INSERT FIG 3.8

Visual Pathways in the Brain What/where hypothesis One path for identifying Temporal lobe lesions in monkeys Can indicate where but not what Another for spatially locating Parietal lobe lesions in monkeys Can indicate what but not where

Visual Pathways: Alternative What/how hypothesis Where something is located in space How do we reach for it?

Theories of Perception Bottom-up theories Parts are identified, put together, and then recognition occurs Top-down theories People actively construct perceptions using information based on expectations

Bottom-Up Processing Theories Bottom-Up Processing Theories Direct perception Template theories Feature-matching theories Recognition-by-components theory

Gibson’s Theory of Direct Perception The information in our sensory receptors is all we need to perceive anything Do not need the aid of complex thought processes to explain perception

Gibson’s Theory of Direct Perception Use texture gradients as cues for depth and distance Allows us to perceive directly the relative proximity or distance of objects INSERT FIG 3.10

Gibson’s Theory of Direct Perception Mirror neurons start firing 30-100 ms after a visual stimulus INSERT VIDEO #18, Mirror Neurons

Template Theories Basics of template theories Template Theories Basics of template theories Multiple templates are held in memory To recognize the incoming stimuli, you compare to templates in memory until a match is found One click to view the words “See Stimuli” and then the demonstration will illustrate the template theory. All images obtained from Microsoft clips. Search memory for a match See stimuli

Template Theories Weakness of theory Problem of imperfect matches Template Theories Weakness of theory Problem of imperfect matches Cannot account for the flexibility of pattern recognition system One click to view the words “See Stimuli”, then the demonstration of the weakness of the theory will occur. All images taken from Microsoft clips. If these clips are being removed, then you can INSERT FIG 3.11 instead Search for match in memory See stimuli No perfect match in memory

Feature-Matching Theories Feature-Matching Theories Recognize objects on the basis of a small number of characteristics (features) Detect specific elements and assemble them into more complex forms Brain cells that respond to specific features such as lines and angles are referred to as “feature detectors”

Pandemonium Model Four kinds of demons Image demons Feature demons Cognitive demons Decision demons INSERT FIG 3.12

Navon (1977) Participants asked what they saw on the Global level Local level Global precedence effect found when letters were closer together. INSERT FIG 3.13 & FIG 3.14 Results depended on whether letters are more widely spaced Participants were faster at identifying local features of the letters

Physiological Evidence for Features Physiological Evidence for Features Hubel & Wiesel (1979) Simple cells detect bars or edges of particular orientation in particular location Complex cells detect bars or edges of particular orientation, exact location abstracted Hypercomplex cells detect particular colors (simple and complex cells), bars, or edges of particular length or moving in a particular direction INSERT FIG 3.15

Recognition-by-Components (RBC) Theory Recognition-by-Components (RBC) Theory Biederman (1987) Describes how 3D images are identified Breaks objects down into geons Objects are identified by geons, relationship between them INSERT FIG 3.16

Evidence for Geons Biederman & Cooper (1991) Evidence for Geons Biederman & Cooper (1991) Used visual priming to demonstrate the existence of geons in a picture naming task Subjects were shown a series of fragmented pictures and were asked to identify the objects BUT there are neurons sensitive to viewpoint-invariant properties Biederman, I., & Cooper, E. E.  (1991).   Priming contour-deleted images:  Evidence for intermediate representations in visual object recognition.  Cognitive Psychology, 23, 393-419.  

Top-Down Processing (Constructive Approach) Top-Down Processing (Constructive Approach) Perception is not automatic from raw stimuli Processing is needed to build perception Top-down processing occurs quickly and involves making inferences, guessing from experience, and basing one perception on another

Evidence for Top-Down Processing Evidence for Top-Down Processing Context effects INSERT FIG 3.9

Configural-Superiority Effect Objects presented in context are easier to recognize than objects presented alone Task: Spot the different stimuli, press button INSERT FIG 3.17

Configural-Superiority Effect Target Composite Pomerantz (1981) Measure reaction time Target alone = 1884 Composite = 749 Target spotted faster in a context!

Which Approach Is Right? Top-down or bottom-up Perhaps a bit of both

Object Perception Viewer-centered representation Object is stored in the perspective seen Store multiple views of object as seen under various conditions Viewpoint dependent process Object-centered representation Object is stored in a way that best represents the object Viewpoint invariant process Realism Link: Why is this important? Understanding how information is coded in our minds helps us to better understand how to create computers that can also perceive objects or to enhance our own recognition abilities.

Object Perception Evidence supports both How to reconcile? Maybe both contribute to object recognition Two ends of a continuum that contribute to object recognition Burgund & Tarr researched this issue

Landmark-Centered Orientation Information is coded by its relation to a well-known or prominent item Consider your college campus What is a prominent item you use to orient yourself on campus?

Gestalt’s View of Perception Basic tenet “The whole is more than a sum of its parts” Law of Prägnanz Individuals organize their experience in as simple, concise, symmetrical, and complete manner as possible

Gestalt’s Principles of Visual Perception INSERT FIG 3.18

Gestalt’s Principles of Visual Perception Figure-ground Organize perceptions by distinguishing between a figure and a background Proximity Elements tend to be grouped together according to their nearness Similarity Items similar in some respect tend to be grouped together INSERT FIG 3.19

Gestalt’s Principles of Visual Perception B Continuity Based on smooth continuity, which is preferred to abrupt changes of direction Closure Items are grouped together if they tend to complete a figure Symmetry Prefer to perceive objects as mirror images A D C Check out Kanizsa’s figure at http://www.psy.ritsumei.ac.jp/~akitaoka/hokan-e.html. It appears as if a white square were in front of four black circles, representing both closure and symmetry.

Pattern Recognition Systems Feature analysis system Recognize parts of objects Assemble parts into wholes Configurational system Recognize larger configurations

Evidence for Separate Systems Tanaka & Farah (1993) Participants studied Faces and names Pictures of homes and home owner’s names At test, given only a piece of face (e.g., nose), whole face, whole home, or a piece of the home (e.g., window) Asked to recall names

Tanaka & Farah (1993) Results People have more difficulty recognizing parts of faces than parts of houses INSERT FIG 3.21

Fusiform Gyrus in Temporal Lobe Implicated in pattern recognition Studies illustrate it is active in facial recognition However, also active if high expertise in any item (birds, cars) recognition Expert individuation hypothesis

Evidence for Separate Systems Prosopagnosia Inability to recognize faces after brain damage Ability to recognize objects is intact Associative agnosia Difficulty with recognizing objects Can recognize faces

Perceptual Constancy Object remains the same even though our sensation of the object changes Size constancy vs. shape constancy INSERT FIG 3.23 & FIG 3.24

Depth Perception The ability to see the world in three dimensions and detect distance Vision only has a two-dimensional view We must interpret the information given to perceive depth We take flat images and create a three-dimensional view Optical illusions demonstrate that this interpretation does not always have to be correct

Monocular Depth Cues Texture gradients Relative size Interposition Grain of item Relative size Bigger is closer Interposition Closer are in front of other objects Teaching Tip: Use a photograph to illustrate cues these cues. Microsoft clips is a good starting point.

Monocular Depth Cues Linear perspective Aerial perspective Parallel lines converge in distance Aerial perspective Images seem blurry farther away Motion parallax Objects get smaller at decreasing speed in distance Teaching Tip: Use a photograph to illustrate cues these cues. Microsoft clips is a good starting point.

Binocular Depth Cues Binocular convergence Binocular disparity Eyes turn inward as object moves toward you; brain uses this information to judge distance Binocular disparity Each eye views a slightly different angle of an object; brain uses this to create a 3D image

Agnosias, Ataxias, & Cognition Demonstrate the modularity of cognition Help us to understand what brain locations are associated with different types of higher-level processing Provide us with a model of how normal processing must work

Deficits in Perception Disruption of the “what” pathway Agnosia Inability to recognize and identify objects or people, despite having knowledge of the characteristics of the objects or people Shows the specialization of our perceptual systems

Deficits in Perception Disruption of the “what” pathway Simultagnosic Normal visual fields, yet act blind Perceives only one stimulus at a time—single word or object

Deficits in Perception Disruption of the “what” pathway Prosopagnosia Inability to recognize faces, including one's own Cannot recognize person from face Knows a face is a face   Can recognize individuals from voice Can recognize objects Can discriminate whether two faces are same or different

Deficits in Perception Disruption of the “how” pathway Optic ataxia Cannot use vision to guide movement Unable to reach for items