Textbook Notes (368,781)
Canada (162,164)
Psychology (3,337)
PSYC 2390 (118)
Lana Trick (87)

Object Perception.docx

17 Pages
Unlock Document

PSYC 2390
Lana Trick

What is Object Perception? 10/22/2012 1:57:00 PM What is Object Perception?  Seeing an object as a whole  We are pretty good at object perception o Sometimes more challenging than others  Accomplishment because there are 126 million rods and cones in each eye o Each rod and cone only knows what they are seeing, only 1 tiny point of light (“pointalistic” representation) o Objects can take up large portion of retina o Sometimes hard to differentiate foreground and background (disrupts object perception)  Different patterns can emerge at different times, differences in organizing information Goals of Object Perception: Object Recognition:  Deciding what goes together to define object  Need to be able to recognize what you see and understand what to do under this context o Basis of intelligent behaviour  learn from experiences so you don‟t make same mistake twice  Needed in order to name things o Sometimes might not be able to name something but would be able to recognize it (sense of familiarity) o Biederman machine  Even a bizarre object that you have never seen, doesn‟t take long to form memory of it  Take visual scene, must perceive object and form representation in your mind o Representation  way of symbolizing info to yourself o When you see elephant in mind, have pattern of activity that “sees” an elephant  Compare this representation it with representation in LTM o Does this match or not? Visual-motor Coordination:  Needed in order to perceive object as whole and then coordinate your body accordingly Why is it difficult to build a perceiving machine?  Stimulus on the receptors is ambiguous o Inverse projection problem  A particular image can be created by many different objects  Humans solve problem by reflecting on past experiences and changing their viewpoints  Objects can be hidden or blurred o Occurs any time one object obscures part of another object o Humans easily understand that even though object is covered it still continues to exist o Humans can also identify objects that are not in focus (while computers perform poorly)  Objects look different at different viewpoints o Images of objects are continually changing depending on angle but we perceive as same object (viewpoint invariance)  Gestalt psychologists  the whole cannot be described merely as the sum of its parts Two Approaches 10/22/2012 1:57:00 PM Two Approaches Top Down Theories:  Use knowledge, memory and expectations to determine the object you see Examples of Top-Down Effects  ABCDEF vs. 10 11 12 13 14 15 o B vs. 13 o Look similar in font o Depending on context, we perceive as B or 13  Can drive interpretation of a object based on priming o What you experience immediately before, can impact what you experience later on o Perceptual set  preparation to see a certain thing  Can have 2 people that have identical experiences but see different things o Your knowledge and beliefs can drive what you see o Auditory perception  hear rock and roll backwards hear worship the devil, those who hear it are highly religious and those who expect to hear worship the devil Disadvantage of Top-Down Theories  Animal that only sees what they expect to see is not evolutionarily adaptive Bottom-Up Theories:  Take info right from retina, no previous knowledge  What you see is what you see Overview of the Stages in Modern Theories 10/22/2012 1:57:00 PM Overview of the Stages in Modern Theories of Perception and How Theories Fit Together Retina  low level vision (pre attentive analyses, spatially parallel)  visual routines (analyses requiring spatial attention)  visual cognition (object recognition) Distinction between 2 Different Types of Process: Spatially Parallel Processes (pre attentive)  Have a process that is looking everywhere in space, at the same time  Doesn‟t matter how many blue objects there are, will always be something looking for red  Tend to just happen whether you want them to or not  Analyses happen on every part of image at same time  Saccades: rapid eye movements, always happen regardless if you are moving your eye, always scanning scene Spatial Serial Process (attentive)  Attentional spotlight or focus on one part of visual image (fixation) o One part of image is getting lots of processing done on it, while others are not o If brain had to deal with all visual information at once, it would become overloaded  Require spatial attention , “zoom lens” of attention  Quite slow  go one area at a time  Why people drive into things they are looking right at o Looking but didn‟t see (weren‟t paying attention, were not attending to what they were looking at) o Texting while driving 16x worse o Attending to one thing, are not attending to something else (don‟t even have to be looking at it) o Conversely, can pay attention to something that you are not looking at (passing ball to someone without looking at them)  Stimulus salience: characteristics that stand out because of physical properties such as colour, brightness, contrast, orientation o Capturing attention this way is bottom up process o Initial fixations are closely associated with high saliency  Knowledge of scene can also impact where we look o More likely to recognize stop sign at intersection than stop sign at straight road  Eye fixations are closely linked to action person was about to take Three Stage Model Overview: Stage 1 (Low level vision – pre attentive)  Spatially parallel  Bottom up processing  Does not require attention Finding Discontinuities in the Image (finding the contours):  Difference between two things that are side by side o Differences in colour (type of light), brightness (amount of light) o Difference defines the edge / line What happens without discontinuities in image?  Ganzfeld o Ping pong balls over eyes  see featureless white or grey in front of you o No contours  If stay in ganzfeld for long period of time, begin to hallucinate little flashes of colour and eyes being to roll around independently, you don‟t know you are doing it (lose track of where your eyes are moving because you have no contours to look at)  As soon as you see contours and have landmarks your eyes are fine o Or by closing your eyes, a white out What happens without change?  Important that positions of contours change  Stabilize image technique o Making sure image stays on exactly same point on retina o Have contact lens, as person moves eye, so does contact lens and therefore image o If object doesn‟t change, image will disappear on you Registering edges (contours) takes time  Masking study o Have stimulus (black dot) that is on for 20 ms  If you see only white afterwards and you see the black dot o What if you see mask afterwards? (black on outside, white on inside and white on inside is same size as black dot stimulus) only one for 20 ms o If time between black dot and mask is less than 100 ms., will see big black dot (visual system added two images together) o If time between 100-200ms, will only see mask (will not add images together) o If time more than 200ms, will see black dot first then mask second  If you don‟t give visuals system enough time to register contours, will do weird things o If too short, will add together o If medium, will totally forget about first and only see second Categorizing the Discontinuities into Types (finding features & feature analysis): How do you know an attribute is a feature?  Treisman o Taking features and asking what kind of feature is this? o Visual search task (“pop out”)  Trying to find a certain target  If target is there press “present”, if not, press “absent”  Manipulate number of items in display (display size is number of objects in the display)  This kind of processing is spatially parallel  Doesn‟t matter how many objects there are, will always find target at same time  Response time (absent and present) is not reflective of amount of objects in the display  Proof for spatially parallel process Different types of features  Colour, line orientation, size Group the Disconnected Contours:  Put them together into higher order groups  Binding: features are combined to create our perception of a coherent object o Experience an integrated perception of object with all components coming together Gestalt Grouping Principles  The whole is different than the sum of its parts  Don‟t see individual elements, see larger units  When see bunch of disconnected elements, we want to group them in image to create larger objects (laws of perceptual organization)  Law of pragnanz / good figure / simplicity: central law, every stimulus pattern is seen in such a way that the resulting structure is as simple as possible  Law of proximity: things that are near each other are grouped together  Law of similarity: similar things appear to be grouped together (size, shape, colour)  Law of good continuation: points that when connected, result in straight or smoothly curving lines are seen as belonging together and the lines tend to be seen in a way to create smoothest path  Principle of common region: elements within the same region of space appear to be grouped together  Law of closure: group to see a closed figure  Principle of uniform connectedness: things that are touching are perceived as single unit (lightness, colour, texture, motion)  Principle of synchrony: visual elements that occur at same time are perceived as belonging together  Law of common fates: things that are moving in same direction appear to be grouped together  Law of familiarity: things that form patterns that are familiar or meaningful are grouped together Some side-effects of grouping  Effortless texture segregation o Can make higher order figures out of contours o Divide image into 2+ surfaces depending o Only occur if parts differ in a feature (colour, line orientation, size)  T and L do not have feature difference (| and _ are component and only difference is spatial orientation therefore harder to see the edge between L‟s and T‟s vs. T‟s and slanted T‟s)  Harder to find an L among T‟s for this reason and takes longer to do so  Formation of illusory contours around object o Kanizsa triangle  Triangle does not exist but visual system “puts one in”  Visual system likes to see continuity  Will group things so it sees continuous lines  Brain perceives image as triangle o Makes contours to look like continuous lines Determining what is figure and what is ground  Figure typically stands out from background  Rubin‟s vase o Grouping phenomenon o Decision about what part of image contour belongs to determines what image you see o Reversible figure-ground  Effects of figure/ground segmentation o Has big effect on what you pay attention to and what you remember o Give participants ambiguous figure (can see either part as figure)  Some chose to make black part figure and some chose white part as figure  Even though same contours, those who chose white as figure, did not see black image even though it was part of the image  What you decide figure is in image is what you remember (if didn‟t see black, won‟t remember black as figure in image)  Things that influence whether something is seen as figure or ground o Tendency to want to make smaller thing the figure  Ba
More Less

Related notes for PSYC 2390

Log In


Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Sign up

Join to view


By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.