Unit 9 Percieving Objects Flashcards
How do we perceive size and how far away an an object is from us
we perceive depth and size by using information from the environment as well as information physical sensations produced in the eyes
How do we perceive a 3D image from 2 dimensional information?
Answer: cue approach
the study of information in the RETINAL IMAGE that provides cues (or clues or hints) regarding the depth of object in a scene
-> we learn how cues relate through experience
What are the three main types of cues that inform us about depth
Oculomotor: cues based on our ability to sense the position of the eye, and tension is our eye muscles (convergence angle)
Monocular: cues based on the visual information available in one eye
Binocular: cues based on the visual information from both eyes
convergence angle (more useful than Acc) - inward movement of the eyes when we look at nearby objects (used by nervous system to calculate distance)
what is accommodation in oculomotor cues
-> the change in shape of the lens, using ciliary muscles (produces tension), that occurs when we focus on objects at various distances
thicker = closer
flatter = further
-> accommodation and convergence => detecting depth information of an object up to ARMS LENGTH
Monocular cues
-> info about depth of object with ONE EYE
includes
-> Accomodation: lens shape
-> pictorial cues: source of depth info in 2D pic
-> Occlusion: when an object in a scene hides another object from view (relative distance)
-> Relative height: object higher in visual field are generally percieved as further (below horizon = true, above horizon = opposite true)
->Relative size: knowing two objects are the same size gives us the impression its further when one is smaller
-> Perspective convergence: one p. drawing affects
->Atmospheric perspective: further = more dust, air, or water
-> texture gradient:equally spaces apart objects, don’t look equal in a perspective picture
-> Shadows: shows us how far something is from a surface/ and how 3d it is
-> motion cues: source of depth info CREATED by movement
-> Motion parallax: when we move by objects, those closer seem to be moving faster than the horizon for example (objects move across retina at different speeds)
-> Deletion and Accretion: when we move some objects are covered from our view and some become uncovered
some cues are more important than others -> depends on distance
Binocular Cues
-> takes into account the differences between the images formed in both the left and right eye
-> each eye detects a slightly different image
right eye closed image of near finger covers image of far finger on left retina
-> from bases of stereoscopic depth perception
2d = monocular cues
3d= add stereoscopic depth perception (recorded using two cameras)
Feature integration theory - what is the binding problem
question such as:
how do we perceive stable, coherent objects in the environment? How do we combine the individual characteristics of an object to perceive a whole, coherent object?
Feature integration theory stages
-> offers answer to binding problem: we have already applied FIT to visual search, and it applies to object perception in a very similar way
stage 1: pre attentive stage (feature search): individual features of the object are detected independently, automatically and in parallel
stage 2: focused attention stage (conjunction): features combine into a coherent perception of an object
Illusory conjunction
Treisman & Schmidt (1982)
-> divided task of report on shape and numbers in two to see if it affect feature collection: it didn’t
bcs according to FIT detecting individual feature does not require focused attention
HOWEVER focused attention (conjunction) stage was disrupted; combining wrong features together (show 2 stages)
Recognition-by-component theory
Biederman’s (1987)
explains how we form coherent perceptions of objects
-> simple geometric shapes AID in recognition: Geons
-> recognising objects by detecting their edges and dividing the object into geons
different set of neons can be organised into different objects
GEONS=VIEWPOINT INVARIENT
Viewpoint invarient
are equally recognisable from any perspective or orientation
-> objects constructed by geons are EASILY RECOGNISABLE from various perspectives
recognitionbycomponent theory explains how we recognise chairs and faces BUT IT DOES NOT SAY how we recognise specific ones
Prototype theory
Rosch’s (1973) -> we recognise objects and place them into caategories by comparing them with an object that best represents the category: prototype
averaged model of all objects we have previously encountered in the category
-> but member of a category vary in the extent to which h they share similarities with the prototype= TYPICALITY
TYPICALITY
high typicality = member have high similarities with the prototype
low typicality = member has low similarity with the prototype
Bosch quantified typicality by presenting participants a series of photos of objects, making them rate to what extent each member represents the category (1very rep to 7 bad rep)
Rosch and mercies 1975 high family resemblance
participants were given several objects and had to list as many of their characteristics as possible -> like a chair> can sit, has legs, supports back (objects could be put into same high family resemblance)
things could be part of the same category but they could be considered low family resemblance like a mirror and a phone
typicality and family resemblance
highly typical members of a category share many characteristics with other members of the category, and less typical members share few characteristics.