Visual Perception Revision Notes PDF

Summary

These notes provide an overview of visual perception, including topics such as object recognition, initial representations, and colour perception. They discuss different models and theories, and explore concepts like sensation and perception.

Full Transcript

Visual perception revision notes Object recognition (Ellis and Young) Initial representations are NOT matched against object recognition units (no link) Hirearchical model- if viewer perception is not achieved, the rest of the model does not work Initial representation (primal sketches)- ability...

Visual perception revision notes Object recognition (Ellis and Young) Initial representations are NOT matched against object recognition units (no link) Hirearchical model- if viewer perception is not achieved, the rest of the model does not work Initial representation (primal sketches)- ability to detect brightness, definition of edges and contours (sensation) Viewer centered representation (21/2 D sketch) edges and contours are assembled to create an object, but only a viewer centered one- e.g a representation from a specific angle, from the observer’s viewpoint, perception of basic shapes, efron test (shape matching) Object centered representation (3D) ability to perceive the object in an atypical view (if you can view the object in a typical view, the “recognition shortcut” can be used, so object recognition units (semantic system and name retrieval) is still possible Object recognition units- ability to name and categorise the object Semantic system- general info about the object, personal facts about objects (LTM) Name retrieval Apperceptive agnosia- impaired perception- impairment in shape perception Associative agnosia- impaired semantics- perception has no meaning Sensation- un interpreted sensory information Perception- interpreted sensory information (making sense of the information) Recognition- identification of a stimulus Marr’s model: 1. Raw primal sketch- intensity and brightness changes 2. Full primal sketch- geometric organization 3. 2 ½ D sketch- ½ to include depth perception 4. 3D sketch- represents the object independently of the observer’s position Aperture problem: 2D motion of an object is ambiguous- only one edge is visible- the receptive field of a single neuron never covers the whole visual field Consequence: the response of a single direction-selective neuron is not sufficient to indicate the motion direction of a stimulus Colour perception Single opponent receptive fields are good at detecting uniform colour surfaces- not good at detecting colour edges Double opponent receptive fields are good at detecting colour boundaries/edges Colours of solid objects- Chromatic colours- reflected wavelengths (the colour we see) Achromatic colours- all wavelengths are reflected Colours of transparent objects- chromatic colours- transmitted wavelengths (the colour we see) Achromatic colours- all wavelengths are transmitted Protanopia- no L cones Deuteranopia- no M cones Tritanopia- no S cones S- short wave length- blue M- medium wave length- yellow and green L- long wavelength- red Transduction- the transformation of the retinal object representation into an electrical signal- happens in both dark and light adapted eyes First step of colour perception- explained by the trichromatic theory (S,M, L cones needed for the full colour experience) Second step colour perception- explained by opponent process theory (opponent neurons respond selectively to red/green blue/yellow and black/white) Colours mix with light (additive) and pigment (subtractive) Psychology of perception (Bottom up approaches- knowledge guided) Psychophysics principles (Fechner): Method of adjustment- the adjustment of light intensity over time in darkness- just being able to perceive the light Method of limits- stimulus is either detected (100%) or not (0%) - single point of stimulus intensity Method of constant stimuli- how likely the stimulus is detected- many more trials- randomised light intensity Webers law- measuring the difference threshold- minimum intensity difference to discriminate two stimuli Wundt- structuralism (study the elements of consciousness) method- introspection (description of sensory experiences) Stevens (subjective magnitude estimation)- difference between perceived and physical intensity- an increase in the perceived stimulus intensity can be larger or smaller than the increase in the measured stimulus intensity Response expansion- perceived stimulus is larger than measured stimulus Response compression- perceived stimulus is smaller than measured stimulus Gibson- perception must be investigated in the natural environment (ecological approach)(bottom up approach) Direct sensation=direct perception Perception through object affordance- objects tell us what they are and what they are used for- cannot explain erroneous perception Perception is error free/direct and always accurate Invariant information: 1. Optic flow pattern 2. Texture gradient 3. Horizon gradient 4. Gravity 5. Straight lines 6. Reflectance of areas What happens when perception goes wrong (Top-down approach- stimulus guided) Gregory- perception is indirect (interpreted) Perception > Sensation Cannot explain why illusions are not learnt (Top-down approach) Gestalt psychologists- the whole is more than the sum of its parts 1. Proximity 2. Similarity 3. Common fate 4. Good continuation 5. Closure 6. Relative size 7. Surroundedness 8. Orientation 9. Symmetry The eye - More rods (120 million) than cones (6 million) - Centre of fovea- only cones peripheral- rods and cones - Rods take longer to adapt to darkness - Light sensitivity- higher in rods than cones - Acuity- better in cones than rods - Spectral sensitivity- rods are more sensitive to shorter wavelengths, cones more sensitive to longer wavelengths Rods are colour blind and cones are colour specific [Purkinje shift- increased sensitivity to short wavelengths in dark adapted eye] Blue-short Green-medium Red-long light waves Photopic vision- cone dominated Scotopic vision- rod dominated Mesopic vision- rod and cone vision together Nearsightedness- concave, farsightedness- convex correction The brain M, P, K ganglion cells Temporal (ventral)- what Parietal (dorsal)- where Double dissociation- damage of different brain areas leads to opposite pattern of impairments Motion perception Corollary discharge theory- considers retinal motion and eye movements- whether real motion has occurred 1. Image displacement signal- moving stimulus + static eye (difference) (1) moving stimulus + moving eye (no difference) (0) must be opposites to get a value of 1 NEW- retinal motion must be present to obtain a value of 1 2. Motor signal- moving eye (1) static eye (0) 3. Corollary discharge signal- when there is a motor signal, there is a CDS Must be a difference of 1 for real motion to have occurred 6 eye muscles Depth cues (Simulus related- monocular cues- work with one eye) - Pictorial cues- we extract 3D information from still 2D images 1. Occlusion- occluded (partially hidden) objects are further away than occluding objects 2. Relative height- objects higher in the visual field are further away than objects lower in the visual field- unless we are on a hill, then objects with their bases closer to the horizon are further away 3. Relative size- when two objects are equal in size, the one that is further away will take up less of the visual field 4. Perspective convergence- converging parts of objects are further away 5. Familiar size- knowledge about the physical size of objects- e.g that kids are smaller than adults- Epstein (1965) experiment: One pound coin is judged to be further away, because it is known to be larger than the 5p coin 6. Atmospheric perspective- objects further away are less sharp and have a blue tint- particles in the atmosphere scatter light (preferentially short wavelengths which softens and colours further away objects) 7. Texture gradient- denser textures indicate further away objects (Gibson’s invariant visual information) 8. Shadows- the more shadow we apply the more 3D depth we create - Motion based- we extract 3D information from animated 2D images 1. Motion parallax- objects that are closer to the viewer appear to move faster than the objects that are further away 2. Deletion & Accretion- an object being covered and uncovered is further away (occlusion in motion) Deletion is the gradual occlusion of a moving object as it passes behind another object. Accretion is the gradual reappearance of a moving object as it emerges from behind another object. Binocular cues (depends on both eyes) 1. Retinal disparity- the retinal image in the left and right eye are slightly shifted because our eyes are a few cm apart Oculomotor cues (state of the eyes) (body-related cues) 1. Accommodation- changing the thickness of the lens to see objects closer and further away e.g ciliary muscles relax when looking at a faraway object 2. Convergence- adjustment of the eyes- looking at a faraway object- eye gaze is parallel Size distance scaling equation- allows us to perceive objects from different distances, as the same size S=K (RxD) S- perceived size R- size of the retinal image D- perceived distance

Use Quizgecko on...
Browser
Browser