PERCEIVING OBJECTSVisual Perception - PDF document

PERCEIVING OBJECTSVisual Perception Aditi Majumder, UCIDifferent Approaches Molecules Circuits &Brain Areas Brain Physiological Approach Individual Groups of Psychophysical Approach Aditi Majumder, UCIGestalt ApproachStructuralism : Perception is created by But this cannot explainApparent Movement Aditi Majumder, UCIApparent Movement Aditi Majumder, UCIIllusory Contours Aditi Majumder, UCIWhole is different from the sum of its parts Aditi Majumder, UCIBasic PhilosophyThe whole is different than the sum of its Six principles defining perceptual How do we combine components to perceive the Is their any basic rules that we use? Aditi Majumder, UCIGestalt Principles of Perceptual Organization Aditi Majumder, UCILaw of Simplicity Aditi Majumder, UCILaw of SimilaritySimilar things appear Aditi Majumder, UCILaw of Good ContinuationPoints that when connected are seen as straight or smoothly together, and the lines a way that they follow Aditi Majumder, UCILaw of Good ContinuationPoints that when connected are seen as straight or smoothly together, and the lines a way that they follow Aditi Majumder, UCILaw of Proximity or NearnessThings close together appear to be grouped Overcomes law of similarity in this example Aditi Majumder, UCILaw of Common FateObjects moving in the same direction appear Aditi Majumder, UCILaw of Familiarity Aditi Majumder, UCIHeuristic and not Algorithm Where does the heuristics come from? Aditi Majumder, UCIPalmer-Irvine Principles of Perceptual Organization Aditi Majumder, UCIQuantitative Measure of Grouping EffectsRepetition Discrimination Task Aditi Majumder, UCIPerceptual SegregationGestalt TheoryReversible figure groundGround is uniform region Separating contours appear to belong to figure Aditi Majumder, UCIFactors Determining Figure and GroundHorizontal or Vertical Aditi Majumder, UCIFactors Determining Figure and Ground Aditi Majumder, UCIModern ResearchPopular beliefLater proved Aditi Majumder, UCIPerceiving Objects Circuits &Brain Areas Brain Physiological Approach Individual Groups of Psychophysical Approach Aditi Majumder, UCIHow Objects are Constructed?Marr’s computational ModelFeature Integration Theory (FIT)Recognition-by-Components Theory (RBC) Aditi Majumder, UCIMarr’s Theory of Object ConstructionComputational Approach Object’s image Identify edgesand primitives Groups primitivesRaw primal sketch2.5-D sketch Perceived 3-D object Aditi Majumder, UCIMarr’s TheoryComputational ApproachAnalysis of light and dark region of retinal imageUsing natural constraints E.g. Illumination edge vs. geometric edgeProcessed to develop a 2.5D sketch Aditi Majumder, UCIFeature Integration Theory (FIT)Focused Attention StageFeatures are combined to perceive the object Aditi Majumder, UCIFIT : Preattentive StagePop-out boundary for detecting featuresDifferent OrientationDifferent Value Aditi Majumder, UCIFIT : Preattentive StageVisual Search Detection TimeConstant with increase in number of distractors if target has pop-out featuresIncreases with increase in number of distractors if target has no pop-out featuresHave to scan each distractor and eliminate Similar to Salience Aditi Majumder, UCIFIT: What makes things pop out?TiltDirection of Illumination Aditi Majumder, UCIFIT : Preattentive StageIndependent featuresNo association with Same observation from Aditi Majumder, UCIFIT : Focused Attention StageAttention is essential for combining features Aditi Majumder, UCIRecognition-by-Components (RBC)Volumetric PrimitivesPrinciple of componential recovery Aditi Majumder, UCIProperties of GeonsView invariance Cannot be mathematically enforced. Not very formal. Limitation of many Aditi Majumder, UCIRBC theorybased on a few basic -Cannot help us detect causes difference Aditi Majumder, UCIComprehensive Model Aditi Majumder, UCIImage Based Stage Aditi Majumder, UCIImage Based StageDifficult Similarity with Marr’s Aditi Majumder, UCIImage Based Stage2D structure of image intensitiesFeatures like edges, blobs, corners etcTwo dimensional Aditi Majumder, UCISurface Based StageVisible surfaces from which light reflect to our Aditi Majumder, UCISurface Based Stage3D surface can be represented by infinite 2D planar elementsDistance from viewerShading (as color or texture)Like a 2D rubber sheet wrapped on the face of the Similarity with Marr’s Aditi Majumder, UCISurface Based Stage2D planes embedded in 3DViewer dependent Aditi Majumder, UCIObject Based StageWe have some 3D Two types of Recognition by Aditi Majumder, UCIObject Based Stage2D planes embedded in 3DVolumesObject dependent Aditi Majumder, UCICategory Based StageHow it helps us surviveWhat about more frames? The temporal domain is not explored that much Aditi Majumder, UCIRelationship to GraphicsOpenGL triangular rendering2D triangle mesh embedded in 3DTriangle is smallest planar 2D elementsVolume RenderingThat is why a view dependent rendering scheme is adoptedSince no object based information, difficulty in handling Aditi Majumder, UCIRole of Intelligence in Object PerceptionAmbiguous Stimulus