Problem Set #2-Problems Flashcards
How does the notion of an accidental viewpoint help explain why we mistakenly see the left-hand scene (below) as containing concentric circles floating in the air, instead of as it really is, at the right?
Accidental viewpoints are highly unlikely. In order to see the scene at the left as a
number of oddly shaped lines painted on the walls (which, of course, it actually is), you would have to assume that you are viewing the scene from an accidental viewpoint where the lines just happen to align into perfect, concentric circles. This is highly
unlikely, so instead your visual system assumes that you really are seeing perfect,
concentric circles floating in space (which is wrong)
The land squid on planet Nexus has three types of cones, with spectral sensitivity
indicated in the plot below. What is the minimum number of wavelengths required to
match any hue that the land squid can perceive? Is there any single wavelength that
looks like white light to the land squid?
1) From the plot, we can see that there are three distinct spectral sensitivities, meaning that the land squid can distinguish between three different wavelengths of light. Therefore, at a minimum, we would need three distinct wavelengths to match any hue that the land squid can perceive.
2) There is a single wavelength, namely 500 nm, that activates all three cone types equally. To the squid, this wavelength will be indistinguishable from white light. (For most
trichromats, including humans, there is no such single wavelength.)
Suppose the
spectral reflectance function of your hand was hard-wired into your visual system at
birth. How could the visual system use this information to achieve colour constancy
If we know the reflectance function of our hand, we can use it as a reference to estimate the amount of energy at each wavelength in a light source. For instance, if the reflectance function of our hand shows that it reflects 50% of light at 500 nm, and we observe that 100 photons of light at 500 nm are reflected from our hand, then we can conclude that the light source actually emits 200 photons at this wavelength. By knowing the spectrum of any light source, we can compensate for changes in the light source and achieve perfect color constancy even when illumination changes.
Sometimes we say that light looks white when it activates all three cone types equally.
Explain whether colour constancy means that we have to revise this claim.
Yes, we have to revise this claim, because colour constancy implies that the colour we
perceive depends both on the cone activations and on the visual system’s estimate of the
spectrum of the scene’s lighting.
Light that doesn’t activate the three cone types equally can sometimes look white. For
example, suppose the visual system estimates that the scene’s lighting has twice as
9
many photons at wavelengths that activate the S cone as at wavelengths that activate L
and M cones. Then, a light that activates the S cones about twice as strongly as the L
and M cones will look white.
Also, light that activates the three cone types equally can sometimes not look white
Colour constancy:
The way we perceive the colour of an object depends on both the colours that activate our eyes and our brain’s understanding of the lighting in the environment.
Accommodation is an absolute depth cue. Why can’t your visual system use
accommodation to determine the true distance to the moon
The eye’s state of accommodation can be used as an absolute depth cue only over this range of distances 0 m to 2 m, . The
distance to the moon is much greater than 2 m, so its distance cannot be estimated from accommodation.
- Suppose your head shrinks overnight, and as a result your eyes are closer together.(Assume that your eyes stay the same size.) If your brain does not compensate for this
change, what will happen to your depth estimates from vergence?
If your head shrinks and your eyes get closer together but your brain doesn’t adjust, then your eyes will see things as if they are farther away than they actually are. This is because your brain uses the distance between your eyes to help you see how far away things are. If your eyes are closer together, your brain will think things are farther away.
Points on the horopter have zero disparity. Does this mean that disparity is not
useful for perceiving the depth of these points?
Disparity is useful for perceiving the depth of these points as well. If a point has zero disparity, that tells us a great deal about its depth, namely that it is on the horopter, and
not in front of or behind it.
Why is disparity not a useful depth cue for perceiving the shape of very distant objects, like faraway mountains or the
moon?
Disparity is not a useful depth cue for perceiving the shape of very distant objects like faraway mountains or the moon because they are too far away for our eyes to see them from different angles. Disparity works by comparing the difference in the images that each eye sees when looking at an object from slightly different angles. But for very distant objects, the difference between the images seen by each eye is so small that it cannot be used to accurately perceive their shape or depth
