Lecture 4 - Depth and visual scene analysis Flashcards
Why is it difficult to design a perceiving images machine?
- Visual input provides ambiguous information about the 3D structure of the world.
- The same 2D retinal image could be produced by an infinite number of 3D objects.
- Image complexity creates difficulty for computers to organsie the visual scene into distinct objects.
How do we perceive depth?
- Humans use a variety of image information to infer depth in a visual scene.
- Monocular/pictural cues
-Binocular cues
What are monocular/pictural cues (work with one eye)?
- relative height
- relative size
- occlusion
- linear perspective
- texture gradient
- motion parallax
-shadows - shading
What are binocular cues (requires both eyes)?
- disparity
What is meant by relative height and size?
- Objects below the horizon and have their bases higher are perceived as distant.
- Of two equal size objects, the more distant one will take up less of your field of view.
- Prior knowledge about relative sizes is needed when judging distance.
What is occlusion?
- Closer objects will occlude (cover) further away ones.
What is meant by linear perspective and texture gradient?
- Parallel lines extending away from the observer converge in the distance. - Texture elements get smaller and more dense with distance.
- Foreshortening (circles become ovals) occurs when the surface is tilted away.
What is motion parallax?
- As we move, more distant objects will glide past us more slowly than nearer objects.
How do shadows and shading affect depth perception?
- Cast shadows create a strong perception of depth.
- Brightness of a surface depends on its oritentation with respect to the light source.
What is binocular disparity?
- the difference in the position of an image on the retinas of each eye, which helps the brain perceive depth.
What is meant by stereoscopic vision?
A slightly different image of the world is received by each eye.
What is disparity?
- This creates differences in image location of an object seen by left and right eyes.
- The size of the disparity depends on the objectʼs depth.
What is the horopter/what does it do?
- Set of points in space that project to correspoding positions in the two retinas (e.g zero disparity).
- Includes the fixation point.
- Objects closer than the horopter have crossed disparities - the image lies further to the right form the right eyeʼs viewpoint than from the left eyeʼs viewpoint.
- Objects further than the horopter have uncrossed disparities - the image lies further to the left from the right eyeʼs perspective.
What is the structuralist approach to perception?
- Wundt pioneered the approach.
- Popular in mid-late 19th century.
- Proposed that perceptuons are simply the sum of ‘atomsʼ of sensation.
- Reductionist approach.
What is the Gestalt School?
- Reaction against structuralism led by three central figures working at Frankfurt University: Wertheimer, Kohler and Koffka.
- Argued that the whole for or configuration is greater than the sum.
What are illusory contours?
- Some images evoke the perception of edges in locations where there is no change in luminance or colour.
- Difficult to explain via the structuralist approach.
What are the Gestalt principles of perceptual organisation?
- Rejected structuralism, proposed principles where elements in an image are grouped to create larger objects.
- Principles are all manifestations of the Law of Pragnanz (ʼgood figureʼ).
How does proximity affect perception?
- Things that are close together tend to group together so we view them as a collective.
How does similarity affect perception?
- Similar things group together.
- This can be based on any basic characteristic e.g. shape, orientation etc.
What is common fate?
- Things that move together group together.
- Motion is a powerful way of segmenting objects from the background.
- Common fate and proximity both required.
What is good continuation?
- Group elements to form smoothly continuing lines rather than abrupt, sharp angles.
- Helps preserve grouping of occluded objects.
What is closure?
- Group elements to form complete figures, despite any incompleteness. - Grouping by closure requires top-down knowledge of forms.
How does symmetry affect perception?
Elements more likely to be formed into group that are balanced or symmetrical.
