1 / 43

Psy280: Perception

This lecture explores the computational problems involved in object recognition, including object constancy, localization, and the role of the ventral and dorsal visual pathways. It also discusses disorders of the ventral pathway, such as agnosia and prosopagnosia, and the neural mechanisms underlying object recognition and consciousness.

carlporter
Download Presentation

Psy280: Perception

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Psy280: Perception Prof. Anderson Department of Psychology Week 5

  2. Visual philosophers • What’s the meaning of life? Not exactly • Computational approach to vision • David Marr (MIT) • What is the purpose of vision? • What are the problems it must solve?

  3. Computational problems in object recognition • What is it? • Object constancy: Variability in sensory information • Retinal position • Viewing position • Occlusion • Lighting, colour

  4. Computational problems in object recognition • Where is it? Where’s Waldo? ?

  5. Two visual cortical pathways • These problems are reflected in the organization of the visual system • Ventral “What” pathway • Inferior longitudinal fasciculus • Dorsal “Where” pathway • Superior longitudinal fasciculus

  6. Landmark Object discrimination Dissociation of what and where in the monkey • Landmark and object discrimination task (Pohl, 1973) • Parietal lobe • Where • Temporal lobe • What

  7. Neuroimaging evidence for “what” and “where: • Attend to change in objects or locations • Objects • Occipito-temporal • Locations • Posterior parietal Same objects Different location

  8. Ventral “what” pathway • V4: Isoluminant color • Lingual, fusiform gyrus • MT (V5): Motion • Middle temporal gyrus

  9. Neuropsychological evidence:Color (V4) and Motion (MT) • Fractionation of perception following cortical lesions • Achromatopsia • Loss of colour vision • Akinetopsia • Loss of motion

  10. Higher-order “what” pathway characteristics • Complex response profile • Dissimilar to V1 • Not simple orientation, colour, motion • Selectivity • Hands, faces etc

  11. Disorders of higher-order ventral visual pathway • Agnosia: “without knowledge” • Visual agnosia: vision w/out knowledge • Modality specific: Restricted to vision • Not a memory disorder • Item can be recognized through other modalities • Touch, sound, smell

  12. Higher-order cortex is highly specialized: Prosopagnosia • Largely specific to faces • Can distinguish between faces and objects • Difficulty in distinguishing between faces • Facial identification Across category Within category

  13. Is there a region of the brain devoted to faces? • Fusiform face area (FFA) • Right middle fusiform gyrus especially responsive to faces relative to other objects FFA

  14. Neural selectivity: Evolution vs experience • FFA and other objects • “Greebles” • Train to recognize individuals • Experts but not novices activate FFA • Potentially not face specific • Reflects both evolution and experience?

  15. Object recognition: Invariance • Recognize object despite differences in: • Size, orientation, viewpoint, lighting, colour, location • Inferotemporal cortex (IT) • Anterior ventral stream • Invariance in neural response

  16. Evidence for constancy: Lateral occipital complex (LOC) • Likely locus of object constancy • Reduction in fMRI response w/ repetition • Invariance • Size, location,viewpoint, illumination, occlusion No effect of occlusion

  17. How does cortex represent all objects? • Specificity coding • Grandmother cells • Every orientation • Every color • Distributed coding • “Coarse” coding across neurons • Combinatorial • Color • 3 cone types • Form • Shape columns in IT • Geons Geons

  18. Ventral stream neurons and consciousness • Binocular rivalry • fMRI evidence: FFA and awareness • FFA turns on when aware of faces • FFA turns off when unaware of faces • Ventral stream representations support consciousness FFA PPA

  19. Dorsal pathway: Action • Double dissociation • Agnosia vs. optic ataxia • Apperceptive Agnosia • Ventral stream damage • Impaired perception • Intact action • Appropriate reaching • grasping • Optic ataxia • Dorsal stream damage • Intact perception • Impaired action

  20. Break! 10 minutes please

  21. How does it all come together? The binding problem • Division of labour: Parallel processing • Colour, shape, motion, depth, location • All in separate regions • How bound together? • Unified perception • Not separate features

  22. Integration (binding) across feature maps • Synthesis requires attention—allows coherence across feature maps: Objects

  23. Attention is mental glue • Allows features to stick together • Without which, perception falls apart • No coherent perception of world

  24. What is attention? • Everybody knows what attention is. It is taking possession by the mind, in clear and vivid form, of one out of what seem several simultaneously possible objectsor trains of thought. Focalization, concentration of consciousness are of its essence. It implies withdrawal from somethings in order to deal effectively with others, and is a condition which has a real opposite in the confused, dazed scatterbrain state …. William James (1890)

  25. What is attention? • “taking possession of the mind” • Control of the focus of attention • “one out of what seem several simultaneously possible objects” • Inability attending to multiple things at once • “It implies withdrawal from somethings in order to deal effectively with others” • “Paying” attention comes with a cost • It has limited capacities that must be shared • “has a real opposite in the confused, dazed scatterbrain state” • Attention is the glue that keeps perception together

  26. Pay attention!

  27. Visual experience: A grand perceptual illusion • Rich and complex? • Sorry folks, its an illusion • We don’t “see” as much as we believe • Don’t notice big changes in our environment • Change blindness • We fill-in our experience to make it coherent • Little persistence of experience from one moment to the next • Seems coherent

  28. Selective attention • Definition • Process relevant and ignore irrelevant • Why do we need selective attention? • Can’t remember/processes everything • Can be independent from eye movements • Helmholtz (1894) • Fixate eyes • Brief flash • Can pick what to perceive • Perception is not fixed • Not just what your eyes do • Perceptual free will • Can choose what to perceive

  29. Attention: Feature Integration Theory Parallel Cereal • Attention needed to glue features together • Feature vs. conjunction visual search (Treisman) • Parallel (“Pop out”) • E.g., Color, orientation • Requires little attention • Serial • Color and orientation • Requires focal attention • Need to move attention around

  30. Attention is required for binding across feature maps • Need attentional beam • Without attention? • Illusory conjunctions • Unbound features • Recombined • Illusory perception • Need attention to glue features together • Otherwise fall apart Report seeing Yellow Square and purple triangle

  31. Perceptual primitives:What makes a “feature”? Molecules of perception? • Perceptual primitives • Building blocks of perception • Cortical feature maps • Luminance, orientation, color, Motion, depth • Higher-order objects • Synthesis of primitives • Objects defined by conjunctions of primitives • Need attention to bond into compounds No pop-out Pop-out Unique primitives Share primitives

  32. Constructivism vs Gestalt approaches to perception • Structuralism • Perception is created by adding elements together • Perception from the bottom up • Like building a house • Gestalt approaches • Top-down perceptual organization • Perceptual inferences: Best Guesses • “Different than sum of its parts” • Not like building a house • Don’t want best guess, want correct answer • Don’t want illusory foundation or beams!

  33. Constructivism vs Gestalt approaches to perception • Perception is much more than what is projected on retina • Structuralism • inputs determine perception • Gestalt • Active role of perceiver • Same input, many different kinds of perception

  34. Gestalt principles of perceptual organization • Best guesses about world • Heuristics • “Rules of thumb” • Visual system is problem solving for us • Visual intelligence • Takes into account probability of occurrence • What is likely vs not Likely Unlikely

  35. Some of the principles/heuristics • Similarity • Good figure/simplicity • Good continuation All have multiple interpretations but we all tend to agree on one

  36. Visual “guessing”: Shape from shading • Sensitivity to perceptual invariance • Light from above, not from below Square recessed Square raised

  37. Figure-ground segregation • What’s an object? • What’s background? • Mental imposition of depth • Occlusion • Infer continuation of background underneath object • Best guess • Unlikely share same contours

  38. Black White Figure-ground segregation • Factors that influence it • Symmetry • Development • Beauty • Meaning • See arrows? • Emotion • Perceptual autism • Pair faces w/ shock • Bias perception toward vase

  39. Chicken or the egg? • Meaning processed before or after figure-ground segregation? • Top-down perceptual organization over-rides initial processing • Only aware of the final products of this process

  40. Substantial feedback onto early visual processing • Higher-order cortex influences lower-order cortex • More feedback than feedforward connections • Has delayed influence • Neural processing is dynamic • Need to revise classic definition of receptive field Greater # of feedback connections applies to early cortex as well

  41. Meaning and perceptual primitives • Common fate • Can perceive objects based on motion • Higher-order meaning influences lower-order perceptual primitives • E.g. motion

  42. Meaning influences perceptual primitives • Structure from motion • Can perceive objects based on motion • Higher-order meaning influences lower-order perceptual primitives • E.g. motion

  43. The End

More Related