1 Perceptual Processes  Introduction Pattern Recognition Pattern Recognition Top-down Processing & Pattern Recognition Top-down Processing & Pattern Recognition Face Perception Face Perception  Attention Divided attention Divided attention Selective attention Selective attention Theories of attention Theories of attention

2 Perception Process that uses our previous knowledge to gather and interpret the stimuli that our senses register

3 Pattern Recognition The identification of a complex arrangement of sensory stimuli

4 Patterns

5 Glory may be fleeting…

6 The Letter Z

7 Theories of Pattern Recognition  Template Matching Theory  Prototype Models  Distinctive Features Model  Recognition by Components Model

8 Template Matching Theory  Compare a new stimulus (e.g. ‘T’ or ‘5’) to a set of specific patterns stored in memory  Stored pattern most closely matching stimulus identifies it.  To work – must be single match  Used in machine recognition

9 Examples of Template Matching Attempts

10 Used in machine recognition Continue here tuesday

11 Problems for Template Matching  Inefficient - large # of stored patterns required  Extremely inflexible  Works only for isolated letters and simple objects

12 Prototype Theories  Store abstract, idealized patterns (or prototypes) in memory  Summary - some aspects of stimulus stored but not others  Matches need not be exact

13 Forming Prototypes Faces-- Faces Animated Version Examine the faces below, which belong to two different categories.

14 Forming Prototypes of Faces

15 Prototypes  Family resemblances (e.g. birds, faces, etc.)  Evidence supporting prototypes  Problems - Vague; not a well-specified theory of pattern recognition

16 Distinctive Features Models  Comparison of stimulus features to a stored list of features  Distinctive features differentiate one pattern from another  Can discriminate stimuli on the basis of a small # of characteristics – features  Assumption: feature identification possible

17 Distinctive Features Models: Evidence  Consistent with physiological research  Psychological Evidence Gibson 1969 Gibson 1969 Neisser 1964 Neisser 1964 Waltz 1975 Waltz 1975 Pritchard 1961 Pritchard 1961

18 Visual Cortex Cell Response

19 Gibson--Distinctive Features

20 Letter Scanning Example First, scan for the letter ‘Z’ in the first column of letter strings. Next, scan for the letter ‘Z’ in the second column of letter strings. Which is easier? Why?

21 Letter Detection Task

22 T Z A How a Distinctive Features Model Might Work:

23 Distinctive Features  Theory must specify how the features are combined/joined  These models deal most easily with fairly simple stimuli -- e.g. letters  Shapes in nature more complex -- e.g. dog, human, car, telephone, etc  What would the features here be?

24 Recognition by Components Model  Irving Biederman (1987, 1990)  Given view of object can be represented as arrangement of basic 3-D shapes (geons)  Geons = derived features or higher level features  In general 3 geons usually sufficient to identify an object

25 Examples of Geons

26 Status of Recognition by Components Theory  Distinctive features theory for 3-D object recognition  Some research consistent with the model; some not

27 Recognition by Components Pro – Biederman found that obscuring vertices impairs objects recognition while obscuring other parts of objects has a lesser effect. Pro – Biederman found that obscuring vertices impairs objects recognition while obscuring other parts of objects has a lesser effect. Which is easiest to recognize as a cup? The left or right? Con – Biederman – Not all natural objects can bedecomposed into geons. What about a shoe? Con – Biederman – Not all natural objects can bedecomposed into geons. What about a shoe?

28 Support for Biederman

29 Summary  Distinctive Features approach currently strongest theory  Perhaps all 3 approaches (distinctive features, prototypes, recognition by components) are correct  Regardless, pattern recognition is too rapid and efficient to be completely explained by these models

30 Two types of Processing  Bottom-up or data-driven processing  Top-down or conceptually driven processing  Theme 5 -- most tasks involve bottom-up and top-down processing

31 Thought Experiment  Assume each letter 5 feature detections involved  Page of text approximately words of 5 letters per word on average  Each page: 5 x 5 x = feature detections  Typical reader 250 words/min reading  6250/60 secs =100 feature detections per second

32 Ambiguous Stimulus -The Man Ran

33 Ambiguous Stimulus - The Cat in the Hat

34 Fido is Drunk

35 Reversible Figure and Ground

36 Word Superiority Effect We can identify a single letter more rapidly and more accurately when it appears in a word than when it appears in a non-word.

37 Word Superiority- Non-word Trial

38 Word Superiority: Word Trial

39 Single Letter ‘K’ vs ‘K’ in a word

40 Word Superiority: Single Letter Trial

41 Word Superiority: Word Trial

42 Altered Sentences in Warren and Warren (1970) Sentence that was presentedWord Heard It was found that the *eel was on the axle It was found that the *eel was on the shoe It was found that the *eel was on the orange It was found that the *eel was on the table wheel heel peel meal *Denotes the replaced sound

43 The Effect of Varying Sentence Frame Context on Interpreting an Ambiguous Stimulus The __________ raised (________) to supplement his income. lion tamer zoo keeper botanist dairy farmer botanist

44 The Influence of Stimulus Features & Sentence Context on Word Identification

45 Attention

46 Definitions of Attention  Concentration of mental resources  Allocation of mental resources

47 Divided Attention

48 Reinitz & Colleagues (1974) Divided Attention Condition Subjects count the dots Full Attention Condition No instruction about dots

49 Proportion of Responses that were “old” for Each of Two Study Conditions and Two Test Conditions (Reinitz & Colleagues, 1994). Study Condition Test Condition Full AttentionDivided Attention Old Face Conjunction Faces

50 Divided Attention & Practice  Hirst, et. al  Spelke, 1976

51 UpsetHotelJudgeEmploymentMapIndulgePencilProblemKeyTerrible

52 Selective Attention

53 Selective Attention (Dichotic Listening Task)  Shadowing  Irrelevant Channel  Cocktail Party Effect - Morray (1959)  Wood and Cowan (1995)  Treisman (1960)

54 Dichotic Listening Task T, 5, H LEFT T 5 H RIGHT S 3 G

55 Cocktail Effect

56 Treisman’s Shadowing Study

57 Stroop Effect

58 Filter Models of Attention

59 Capacity Model of Attention

60 Diagnostic Criteria for Automatic Processes

61 Cerebral Cortex & Attention