PSY 310: Sensory and Perceptual Processes 1

Transcription

1 Size perception PSY 310 Greg Francis Lecture 22 Why the cars look like toys. Our visual system is useful for identifying the properties of objects in the world Surface (color, texture) Location (depth) Size In this class, we often mention visual angle as a way of measuring the size of a stimulus But that s not the same thing as the size of an object It would be good to have size constancy where the perceived size of an object does not change with distance We don t quite have that, but we are pretty close Visual angle How do we describe the size of visual stimuli? A larger image that is further way is exactly the same on the back of the eye! The size of the image on the retina depends on the distance of the object We can compensate for distance S = K(R x D) S -> perceived size R -> size of the retinal image D -> perceived distance of the object K -> a constant to keep units valid Consider a snow man 15 feet away and 5 feet tall It produces a retinal image that has a visual angle of θ =18.9 The perceived size of the snowman would be S= K(R x D) S= K(18.9 degrees x 15 feet) S= K (283.8) θ 5 feet If K = degrees -1 S= 5 feet 15 feet Nothing interesting here, I just set the K term to give me the actual height of the snowman 1

2 Consider a snow man 25 feet away and 5 feet tall It produces a smaller retinal image that has a visual angle of θ =11.4 The perceived size of the snowman would be S= K(RxD) S= K(11.4 degrees x 25 feet) S= K (285) θ 5 feet If we keep K = degrees -1 S= feet Small difference is due to rounding errors in the calculation 25 feet We always get the same value for the size! Emmert s Law The size-distance scaling idea explains an odd thing about afterimages They change apparent size depending on the depth of the surface you look at Color afterimage Emmert s Law The afterimage exists only on the retina Or for neural circuits that represent things in retinal coordinates So, it always has a fixed retinal size R = 23 degrees As you look at different places, the perceived depth changes On your desk D = 1 foot S = degrees -1 (23 degrees x 1 foot) = feet On the screen D = 60 feet S = degrees -1 (23 degrees x 60 feet) = feet What happens if you do not have any estimate of perceived depth? Then the perceived size is related only to the visual angle of the image As if the distance was constant Then you cannot accurately judge perceived size The sun and the moon appear to be about the same size 0.5 degrees No perceived depth 2

3 In reality, the sun and moon are very different in size And distance Sun is 93 million miles away Moon is 245,000 miles away No perceived depth Poor depth There is a similar effect when looking from the top of a tall building John Hancock building in Chicago We don t have a good estimate of depth (we estimate it to be shorter than it really is) You don t have to know the distance of something for it to contribute to your size percept It is not a conscious calculation Your visual system does it automatically Like it does for accommodation or convergence of the eyes This can lead to some illusions The horizontal lines are the same size Easy to prove Size illusion CogLab version CogLab version Method of constant stimuli Right line always the same Left line varies from trial to trial The line with wings was always 100 pixels If there was no illusion you would expect that proportion of 0.5 would be at Size of line 100 pixels 3

4 CogLab version The actual pixel size that leads to 0.5 proportion reports indicates the perceived length of the line with wings Why the illusion? The outward wings indicate the line in between is further in depth Automatic process by your visual system Why the illusion? Ponzo illusion If two lines have the same retinal size and one is further away A similar effect explains the size illusion here The further away line must have a larger physical size Illusions Other effects Usually, applying this idea does not lead to an illusion The three figures have the same retinal size, but different perceived sizes Perceived size is intimately tied up with perceived depth, but it s not the only issue Suppose you are a bomber in an aircraft. Your mission is to blow up fuel tanks. You have one bomb left. Intelligence has told you that the center tanks are full and the others are empty. Which one do you blow up? 4

5 Prof. Greg Francis Ebbinghaus illusion Jastrow Illusion The (retinal) size of surrounding objects affects the perceived size of an object Which object looks to be larger? Jastrow Illusion Jastrow Illusion Which object looks to be larger? It involves which part of an object is compared to which part of the other object Which of these drawings shows the moon sized properly? The moon seems to change size Very large when low on the horizon 5

6 The moon seems to change size Smaller when up in the sky In reality, the moon is always the same, and so is it s retinal image Actually there are small effects of atmosphere and the radius of the earth (extra distance) But these make the retinal image of the moon smaller on the horizon The sun and other objects show the same kind of size illusion Explanation Explanation When looking at the moon in an empty sky, there are no cues to the distance of the moon No disparity No occlusion No linear perspective We default to a standard Eye convergence Which contributes to perceived size When the moon is on the horizon there are some cues to distance of the horizon elements Linear perspective Occlusion Atmospheric perspective Relative size The moon is behind all objects So it must be further than them Which leads to certain eye convergence Possibly also to a further distance So the moon must be bigger! Tricky Other interesting size illusions There s still no real agreement on the reason for the moon illusion In particular, the horizon moon often seems closer than the moon in the sky You would expect the opposite There may be multiple effects going on Ponzo illusion Ebbinghaus illusion The two half discs are identical in size and shape 6

7 Other interesting size illusions Conclusions Vertical horizontal illusion We do a pretty good job overall Emmert s law Illusions Next time Review for exam 2. Bring your questions. Then Exam 2 on Friday 7