PSYCH 3420 Cumulative Review
Why does CIE use 'imaginary' primaries? What is represented by the points outside of the CIE chart?
CIE uses 'imaginary' primaries XYZ so that all visible colors can be described by positive combinations of these primaries. The points outside the CIE chart represents these 'imaginary' colors, which can still be represented mathematically by the primaries, but don't actually exist. This is preferred to the alternative of using 'negative' light. If you choose your axes to be within the set of visible colors, then in order to describe all visible colors, you would need to use negative amounts of the primaries. Equation: AtLt = A1L1 + A2L2 + A3L3
182. What is change blindness? Provide an example. What does this say about the amount of information we code about the world or a stranger that walks up to us?
Change blindness is not noticing changes in our visual field because we only focus on a certain amount of the environment at a time. In one experiment, students asked strangers for directions. Someone walked by carrying a large object, like a door, and the student swapped places with someone else. Most times, the stranger was not aware that they were talking to someone different from before. This is also why people don't notice the gorilla in the famous video of the basketball players - since viewers are asked to count the number of passes, they do not notice much else in the scene.
Describe the difference between closed loop, open loop, and partially closed loop displays.
Closed loop: Your behavior affects what you see Open loop: Your behavior does not affect what you see Partially closed loop: Not linked to natural motion, but still affects what you see (i.e. scrolling through a website with parallax, or Google street view which jumps from point to point and doesn't move smoothly/naturally)
To what extent is color symbolism universal across the world?
Color symbolism is used anywhere from advertising to deciding what color to paint your kitchen. There is some universal color symbolism, but most symbolism stems from a particular culture. Examples: Purple - symbol of death/crucifixion in Catholic Europe Black - death, or class and wealth (i.e. limos and suits) White - purity/cleanliness (toilets, wedding dresses) or mourning in China Red - Blood, war, sexual connotation (red light district), stimulating products (Coke, Marlboro)
What is forced perspective? How does it apply to the filming of Frodo and Gandalf together on the cart in Lord of the Rings? Why does motion of the camera make this difficult? What is the solution?
Forced perspective is an illusion in photography and videography that makes objects look like they are in different locations and different sizes than they really are. This is used in Lord of the Rings to make Frodo look small by placing Elijah Wood farther back in the scene than Ian McKellen. The cameras are placed so that they look as though they are interacting side-by-side as opposed to different distances from the camera. This is difficult with camera motion because the props are often built to be at the correct scale for each actor and to be seen at a specific angle. The solution is to put the actors and props on tracks that move as the camera moves.
168. What is so remarkable about the boy that 'clicks' video? What skills did he have?
He became blind around 3 years old from eye cancer. He learned to navigate using echolocation - clicking with his tongue and learning where things were by how the sound waves came back to him. Out of all of the people who learned to do this, he was by far the most successful/talented at it. Fact check- his echolocation abilities actually developed in his visual cortex?
170. Describe three examples of augmented or mixed reality and include an application for each.
1. Google Maps using video feed from your phone. You can hold up your phone camera to the street and it will overlay information on the image your camera sees. It can tell you where to turn based on what street you are on, how far you are from your destination, etc. (MR) 2. Heads-up display while driving. Having a HUD on your car to tell you your speed, location, directions, etc. This can be very overwhelming and possibly more dangerous since you are being bombarded with information and can't focus on the road (AR) 3. Glasses to help in manufacturing. People in manufacturing can wear goggles that will overlay instructions onto their line of sight. For example, someone putting together parts of a car could be shown which pieces fit where and how they need to be connected. (AR) 4. Gaming - Pokemon Go uses MR to fix animations into a video stream of the real world.
172. Other than gaming, describe five current applications of virtual reality.
1. Military: Allows troops to be trained in different combat scenarios and in simulators (like flight simulators for pilots). This is cheaper and safer than carrying out practice missions, and can be redone easily for more practice. 2. Educational: People can take VR tours of museums or wonders of the world without having to spend the time and money it takes to travel there. 3. Medical/Therapy: Patients in a lot of pain (physical and emotional) have found that exploring a VR world while in therapy has helped reduce their pain. 4. Sports: You can view sports games (or concerts, or other events), as though you are actually at the game, and have front-row seats. 5. Practice for public speaking: You can give a speech to a virtual audience to help practice. The VR system can record what you are saying and track your movements to provide you feedback at the end.
192. Provide 3 examples that demonstrate that the visual system is making assumptions about what is likely in the environment. Include examples on the relation between 2D and 3D structure.
1. People assume that if lines come together in 2D space, they must also come together in 3D space. This helps create a lot of the illusions in Lord of the Rings when Frodo and Gandalf interacted with two separate objects (i.e. tables) that looked like one object because of the camera angle. 2. People assume parallel lines in 2D are also parallel in 3D. This can be shown in the illusion of the balls rolling 'uphill'. The camera angle makes the ramps look like the balls are falling up because the stands holding them up look parallel. In reality, the middle stands are angled away from the camera so that the part of the ramp that looks the highest is actually the lowest. 3. People assume angles are 90 degrees wherever possible. This is why the necker cube illusion is confusing
208. Describe two advantages of using animated images over static images (W)
1. Teaching and describing physical movements is much better done with animations than static images. When teaching tennis, for example, a short video of how to hit the ball with the racket is much more effective than a few still images, even if they have arrows showing direction of motion. 2. You can use much simpler graphics if you include some motion. For example, you can indicate pressing a button with a basic rectangle or circle moving back and forth. You would need more detail in the image to show the same task in a static image. 3. Static images of faces can't express a full range of nuanced emotion because this depends on the way the muscles in the face are moving, not just a still snapshot. An animation of a face making an expression could give you a wider range of emotions than still images.
Provide two examples of art that take advantage of the acuity limitations in the visual periphery (Livingstone)
1. The Mona Lisa smile - when you are looking directly at the Mona Lisa's mouth, it does not look like a genuine happy smile. When you look at other parts of the painting, so that her mouth is in your periphery where there is lower acuity, the smile looks genuine, making the painting look mysterious. 2. Rue Montorgueil by Monet - This painting is very spatially imprecise, with large brush strokes and jumbled details. This gives it the effect of motion, and being a snapshot in time, because the imprecision causes you to perceive it/complete the detail slightly differently every time. The style reflects the way our periphery works.
215. Describe four things that you learned from the Ware text regarding techniques that produce more effective displays.
1. The optimal size for recognizing a visual object is 4-6 degrees of visual angle. Designers should take into account the display size and observer distance to design their graphics. 2. It takes a lot of attention to hold 3-4 objects in visual working memory. A simple display could be the best choice, especially for marketing, because one attention-grabbing object is more memorable than a busy display, which can be overwhelming. 3. You can add anchors to charts and graphs to help create a sense of relationship between different points. Even though this is a non-natural sort of design element, because we are sensitive to lines it can help us get a better understanding of the chart. 4. Use lines to show relationships between different objects (i.e. on a flow chart) 5. Sensory symbols (i.e. drawing of a dog) may be better than arbitrary symbols (i.e. the word 'dog') in some cases because they are processed in lower levels and may be more universal. 6. Dynamic displays (videos) may be more effective than static images depending on the context. In teaching, especially teaching tasks requiring motion, a video would be easier to understand than a cartoon or even photographs showing the different steps.
193. Describe four "rules of design" (according to Donald Norman). Provide an example of each.
1. Visibility and feedback - You should be able to see the object of importance (ex. white text on a white background is not visible) 2. Natural mapping - 'if it needs a sign, it's probably a bad design' (ex. a push door that looks like a pull doesn't have natural mapping) 3. Constraints - restricting the interactions that can take place (ex. making a USB one-sided to force users to put it in the right way) 4. Design for error - things should be designed so they are used properly. Take into account the errors the user is likely to make. (ex. spikes for pigeons in NYC, anti-animal cruelty club shirt, where 'anti' is hidden by the hood is bad design)
214. Describe an example of a 'visual thinking algorithm'. (W)
1. Visual queries are a sub-algorithm of other algorithms. The steps are: - identify problem components and create a query pattern to look for - low-level visual system is tuned to be sensitive to the query pattern - an eye movement is made to the next most likely target location based on the scene gist - search targets are processed within the fixational area - repeat 3-5 as needed 2. Pathfinding on a map or diagram - Conduct a visual search to find nodes representing start and end points - Mark the start and end nodes in a mental map - Fixate on the start symbol -- extract patterns of lines extending in the general right direction -- mentally mark the end point symbol of the best candidate line - Repeat from 3 using a new starting point until the destination point is located - Push information concerning the path to the logical propositional store - Repeat from 3 to find alternative candidate paths, avoiding paths you've already seen
Provide two lines of evidence that we are especially sensitive to biological motion.
1. We are very good at determining the sex, age, weight, and even mood of a point light walker 2. Infants at 4 months old can differentiate biological and nonbiological motion 3. We are extremely sensitive to errors in biological motion
216. In terms of visual resolution, how close is the Oculus Rift to visual reality? What resolution is required?
100 pixels/degree is generally accepted as a good enough resolution to give us real-life quality. The Oculus Rift is currently at 8-18 pixels/degree, depending on the model. This is nowhere near realistic eye resolution. The Oculus Rift also only provides about 110 degrees field of view, which is not close enough to our 180+ natural field of view. It causes a strange effect when objects disappear in the periphery of the Oculus Rift.
207. Describe the role of canonical views with object recognition. (W)
A canonical view is a view from which an object is most easily recognized, because it is how we typically see it. For example, we would be faster at recognizing a car if we are looking at it from the side than from the bottom or top. In general, objects are identified faster when they are upright than when they are on their side or upside-down, but scaling the size of the object doesn't seem to affect recognition speed. This also relates to prototypes, because prototypes of objects are from canonical views. Some categories do not have canonical views, because they are so broad (ex. games can mean sports, board games, video games, etc) In visualization, especially in infographics, canonical visuals can be used to represent a group (i.e. a square with a triangle on top is a house, a cartoon hamburger may represent food)
Dichromat
A dichromat is someone who only has 2 types of cones, as opposed to the standard 3
What is a scale composite? Provide an example.
A scale composite is where each actor is shot separately, and the two pieces of film are put together later. This is used with blue screen so that they can be shot at different distances or one scaled down, and still be composited together. Used to make the hobbits look smaller than the elves.
210. What is scene gist? How long does it take to recognize the gist? (W)
A scene gist is when our brain recognizes the type of scene we are seeing (indoor, landscape, city, etc) so the visual system is primed to employ strategies for finding interesting stimuli in that scene type. For example, if we are in a city where there are lots of vertical lines, maybe straight vertical lines are less likely to be important than in an open field, where a straight vertical line could indicate the only building around. It takes about 100 ms to recognize a scene gist.
Shadowing
A task in which a participant has to repeat words they hear as they are being presented
211. What is meant by O'Regan's comment that the world has its own memory? How is this relevant to change blindness? (W)
His comment refers to the fact that we have a very detailed representation of the world not because we have a detailed internal model of all of these details, but because we can focus on them whenever we want. Change blindness shows us that we really don't have a very large visual working memory, which is why we can go without noticing even some very big details (like a gorilla walking through a basketball game). People are often surprised by change blindness because we think we have a great detailed model of the world. In reality, we only think that because we can focus on different details at any time at will, not because it is all stored in memory. We assume that the world around us is fairly consistent. Change blindness takes advantage of this by changing something that we assume is consistent, so we do not even notice.
Tritanope
Absent blue-sensing cone
Deuteranope
Absent green-sensing cone
Protanope
Absent red-sensing cone
What is display efficiency (DE)? What is visual efficiency (VE)? Why is DE never 100% with current technology? (Ware, p56)
According to Ware, Display Efficiency is the ratio of the uniquely stimulated brain pixels to the number of actual screen pixels. Since sometimes multiple brain pixels are triggered by the same signal when looking at a screen, those are counted as redundant and not included in the calculation. If there is a perfect one-to-one match, we would have a DE of 100%. This is not realistic, because screen pixels are uniformly distributed and brain pixels are not. Visual Efficiency is the ratio of the number of uniquely stimulated brain pixels to the total number of brain pixels. It tells us what fraction of our brain pixels are getting unique information.
Spectrum locus
All the colors that we can perceive (~370 to 730 nm)
Describe a color matching experiment that proves that the entire color space is 3-dimensional (can be described in terms of 3 variables).
All visible colors can be described using the following equation as long as all three of the primaries are not co-planar: AtLt = A1L1 + A2L2 + A3L3. If our primaries are all real lights, we will sometimes need 'negative light' to describe them all. If we use 'imaginary primaries', we can use the primaries to describe more than just the visible colors Experiment: That equation will always hold, even for negative colors, even if it is not physically possible to create the color. Just add the negative to both sides of the equation to make it positive
What is amblyopia? About what percentage of the population has it? Why is this relevant to designers of 3D displays?
Amblyopia (lazy eye) has two forms: 1. Strambismic amblyopia/squint: misalignment of the eyes 2. Anistropic amblyopia: one eye out of focus People with amblyopia (about 6% of the population) don't have good binocular vision, and are very dominant in one eye. Designers of 3D displays have to realize that this 6% of the population will not be able to perceive their 3D effects if they rely on binocular disparity. They should use monocular depth cues to allow people with amblyopia to see the effects
What is the difference between the standard anaglyph and Dolby 3D? What are the advantages of Dolby 3D?
An anaglyph is a standard, simple method of creating a 3D effect using a blue filter over one eye and a red filter over the other eye. The display has two images, one red and one blue, that are slightly offset. The glasses filters allow one image into one eye and not the other, creating parallax. In Dolby 3D, each eye is covered by a 3-band notch filter that allows multiple wavelengths through, which are slightly offset for each eye. Advantages of Dolby 3D: - Can get a wider range of colors - There is little ghosting, while anaglyphs have a lot
What is anamorphic art?
Anamorphic art is art that seems stretched out, but when seen at a specific angle the correct image can be seen. For example, the painting 'The Ambassadors' has a skull that can be seen properly by standing to the side of the painting.
179. What is anchoring? Provide an example. In what ways is this an example of subliminal perception?
Anchoring is a bias in your decision making caused by some other source of information (the anchor), whether or not that anchor is relevant to the decision you are making. For example, seeing a photo of a football player with either a 54 or a 94 on his jersey caused people to answer a following question differently. Participants in an experiment were shown one of the two images and then asked to guess the percentage of African nations in the UN. People who saw the 54 guessed lower numbers, and people who saw 94 guessed higher numbers. Anchoring is kind of subliminal perception. It is not below your detection threshold, so you may be aware of the stimulus. The subliminal part is that this may subconsciously influence you, which you are not aware of.
183. Describe two ways that identification of an object based on a single feature differs from the identification based on a conjunction of features?
Identifying an object when only a single feature differs from the rest: all rectangles are one color but only one is horizontal while the rest are vertical Identifying an object by a conjunction of features: all rectangles are either red or green randomly, and vertical or horizontal randomly, but only one is both red and horizontal Differences: 1. Searching task: single feature: parallel + fast, conjunction of features: serial + slow 2. If you have 2 regions of texture and they differ by a conjunction of features they are hard to differentiate. If they differ by a single feature it is easier 3. illusory conjunctions - when a conjunction of shapes and colors are flashed before you, you will tend to mix them up. For example, seeing a blue heart and a red oval, you might say you saw a red heart because you combine the features incorrectly. This can be a big problem in eyewitness testimony.
If R and G are fixed at 100, what colors can be produced by varying the amount of B?
If R and G are fixed at 100 and B is 0, then the resulting color is orange. Adding more blue (up to the max of 100), would create a range of values of orange between a very pure orange and white.
What is the evidence that color names are not culturally determined? What did Berlin and Kay show about the evolution of color names?
If you have a society with color concepts, you can predict what colors they will develop names for first. Berlin and Kay showed that there is a common pattern in color names. Languages with only two color names only had black and white. Languages with 3 colors added red, 4 colors added green or yellow, etc. Most languages have at least black and white color names.
Simultaneous contrast
If you see an image that is tinted a certain colors, the complement of the color will appear to be stronger than it actually is. - a green-tinged picture of strawberries will still appear to be red - Casting a shadow in white light on a blue surface will make the shadow look yellow A gray color surrounded by a color will appear to be the opponent color
Illusory conjunctions
Illusory conjunctions occur when people combine features of two objects into one. For example, when quickly shown a slide with different colored shapes, people often said they saw a red heart (because hearts are usually red) when they really saw a red oval and a blue heart.
What is the difference between LUV (UVW) and xyz color space?
In XYZ color space (like the CIE color chart), the colors are spread out based on their linear combination of three orthogonal axes. In LUV space, the colors are rearranged based on McAdam Ellipses, so that there is equal discriminability throughout the color space. The McAdams Ellipses, and LUV space, take into account human sensitivity to different wavelengths, and adjust the color space so we are equally sensitive to the changes in the color space.
205. Describe the basics of Biederman's 'geon theory'. (W)
In a neural network model, the first two layers detect edges, and then vertices, axes and blobs respectively. The third layer extracts 3D geon attributes, which are primitives like cones, cylinders, and boxes. The next layer describes how these geons interact with each other (location, angle, etc)
What is the difference between additive and subtractive color mixture? Provide an example of each.
In additive color mixture is done by combining multiple sets of wavelengths to create a color. An example of additive color mixture is if you have three spotlights with different colored filters on them and you overlap the light beams. Subtractive color mixture is done by removing wavelengths from a light in order to create a color. CMYK printing is a subtractive technique because the combinations of cyan, magenta and yellow absorb some wavelengths (removing them from the light) and reflect what is left over.
Before the Renaissance, what evidence suggests that artists did not have a solid grasp of linear perspective?
In linear perspective, all parallel lines converge onto a vanishing point. Before the Renaissance, images did not follow this rule.
176. Prism adaptation is local and active. Explain.
In prism adaptation, participants wear goggles with prisms that distort their view. The prisms may shift their view over to one side, elongate it, or flip it upside-down, as an example. As people do stuff while wearing the goggles, they adapt to their new visual perception. Prism adaptation is local because it will only affect the body parts you are using while seeing through the goggles. For example, if you are playing catch with your right hand while wearing the goggles, your right hand will get better at figuring out where the ball will be. If you suddenly switch to your left, hand, you will find it has not adapted and can't catch as well as it normally would. The same is true when you remove the goggles. You will find that your left hand has no problem catching, since it has not adapted to the prism view, while your right hand will need to re-adapt. Prism adaptation is active because it requires you to interact with the environment. For example, your hands will not adapt to catching and throwing with your new distorted vision unless they actively practice catching and throwing.
178. What does the Kunst-Wilson study say about the possibility of subliminal perception? What did they find?
In the Kunst-Wilson study, participants were quickly flashed polygons of different shapes, long enough to detect but not recognize them. They were then shown pairs of polygons (one they had just seen, and one they hadn't), and asked 1) if they had seen one, 2) which one they liked better, and 3) their confidence. They found that participants did not do great at recognizing the ones they had been shown, but did tend to like ones they had seen about 65% of the time. This suggests the mere exposure effect - simply seeing something, even if you are not aware of it, can impact how much you like it.
Explain why Johnny Lee's demonstration of depth on the Wii is so powerful.
It is a closed-loop display - the motion parallax remains correct as you move around the room
What are the advantages of filming at 60Hz or better? How does this improve the appearance over traditional techniques? What are the disadvantages?
It seems that the higher the frame rate, the smoother the resulting video will be. At 60Hz and above, the spacing between frames becomes less perceptible, so the frames look less jerky. The disadvantage of filming at higher frequencies is that it is memory-intensive to store that much more data
What is Livingstone's theory regarding the origin of the Mona Lisa elusive smile?
Livingstone's theory is that the Mona Lisa smile looks different depending on where you are looking in the painting. When her mouth is in your periphery and you can't make out the sharp details, it seems more cheerful, but when you look at it directly and get more detail with your foveal vision, it looks less like an authentic smile and looks more coy or mysterious.
Can I match a mixture of 585 nm and 500 nm (50% of each) with a color on my computer screen? What would be the RGB output?
Yes, the result would be a green with RGB values approximately [0, 90, 10]. Taking the point on the CIE chart halfway between 585 and 500 nm gives you a green on the edge of the RGB triangle . This value of green is about 90% green and 10% blue.
Semantic priming
You are faster at identifying things if they are related to a word you were just primed with.
How many colors can I produce with an 8 bit look up table and 6 bit DACs?
[Check answer] The lookup table can describe 2^8 = 256 possible colors, and the DACs can create 2^6 x 2^6 x 2^6=2^18 =262k colors. The lookup table is the limiting factor, so you can only produce 256 colors. Abstracted: 2^(bits in look up table) out of 2^(# of DACs * bit depth of DAC)
175. What are the limits of perceptual adaptation? a. To what sorts of changes can we adapt? b. What adapts? c. How does the Held and Hein study with kittens relate to prism adaptation?
a) We can adapt to distortions in: - Orientation - Spatial Frequency - Light - Color - Motion b) We adapt: The body part in use adapts, *not* the visual systems - locally to the body part being used (i.e. one hand but not the other) - adaptation is active - requires some sort of movement - adaptation is smooth - cannot adapt to local discontinuities c) In Held and Hein's experiment, two kittens were placed in a carousel. One could move and walk around while the other was passive (pushed around in a box). The passive kitten developed perceptual-motor problems. This shows that adaptation is active; you need to be interacting with the environment physically in order to adapt to the world. In the prism experiment, you will not adapt to your distorted vision unless you actively move around and try to reach for objects around you.
212. What is 'inattentional blindness'? (W)
failure to recognize some stimulus which is in your field of view because you are focusing on something else so intently
171. Describe the basic components required for Oculus Rift and the Magic Leap headset. Briefly describe the perceptual experience of each. What problems are you likely to encounter?
*Oculus Rift = VR* VR Components: The entire world is simulated 1. Head-mounted display 2. Method of tracking head position 3. Interactive input - the view input changes based on the location of the head. In MR this requires some sort of model of the real world. 4. Optional: Means of interacting with the environment *Magic Leap = MR* MR Components Parts of the world are simulated 1. Camera to view the world 2. Ability of software to anchor items to the real world 3. For head-mounted displays, image must be in focus, ideally this is based on where you are looking in the display. Experience: In VR (Oculus Rift), you can only see the virtual world around you. In MR (Magic Leap), you can see virtual objects (e.g. virtual creatures) moving around in the real world through the goggles Problems: - Field of View is not as big as we are used to, so objects may disappear as they go into our periphery - Resolution is not as good as we would like, making things look less realistic - Focal length is usually fixed, even if the system has 2-3 possible focal lengths. Young people will try to accommodate for the distance of an object, which may not be the same as the system's focal distance. This can give people a headache. - (MR) Objects may be clipped as you get too close to them to avoid some accommodation issues, but this clipping looks bad
Describe an example of how 'point of view' was used by Mantegna.
- Painting of Jesus from the feet to bring the observer up close to the dead body - Painting on the ceiling of figures looking down at you
What technologies use additive color mixture? Which use subtractive color mixture?
- additive color mixture: film/projection/any digital media that emits light - subtractive color mixture: hard copy media (painting, photographs etc)
180. What is the relationship between prototypes and face recognition?
Prototypes are the average of a population and the average of the feature positions (e.g. the 'average' face). When we recognize faces and other objects, we are often comparing the object to its variation from the prototype we have in our mind of what the average face would look like. This is why we are so good at recognizing caricatures, which exaggerate the unique features of a person.
169. Provide two examples of sensory substitution that allow the blind to see.
1. Auditory: Headphones that play you the frequencies of the image in front of you, from left to right. High frequencies correspond to brightness near the top of the image and low frequencies correspond to brightness near the bottom of the image. 2. Tactile: Electrodes placed on the tongue. They are stimulated in different ways based on the image in front of the person.
Describe four color phenomena that are explained by opponent color theory that are not explained by trichromatic color theory.
1. Color naming - there's no such thing as 'greenish red' or 'yellowish blues' - those are opponent colors. There are color-opponent ganglion cells which are red/green and blue/yellow as well. 2. Color Afterimages - If you stare at a colored image for a while, then look at a white, neutral surface, then you will see the opponent colors from the image you just looked at. The McCollough effect is an example - it is an aftereffect of staring at colored gratings. The colors of the gratings become associated with the grating orientations. 3. Simultaneous color contrast - color constancy - An image that is tinted appears to contain its contrast color (i.e. a shadow on a yellow background looks blue) 4. Color naming for anomylous vision - someone missing a cone can delete an opponent mechanism. For example, someone missing the red cone, the red-green opponent mechanism disappears People who are color blind in one eye only have shown that the entire opponent mechanism is missing in the color blind eye, not just one color.
How does one make an Ishihara color test chart that finds protanopes and distinguishes them from deuteranopes?
1. Find two colors along a confusion line for either protanopes or deuteranopes 2. Using spots, make a number with one color surrounded by a background of the other color. 3. Randomize the intensities and sizes of the spots 4. Make a number of lower contrast that will be visible to the subject if they can't see the color difference (optional)
191. Describe the difference between 2D, 2.5D, and 3D representations. How can Escher's paintings be explained in terms of this difference?
2D = no depth 2.5D = simulated depth, you interpret it in 3D. Our perception is 2.5D because we see a 2D image but interpret it in 3D because of depth cues 3D = has depth Escher's paintings played with that. They are drawn in 2D, but has depth cues within it, and you interpret it in 3D. However, when you try to figure out what the actual 3D structure is, you find it is not so trivial, and it is difficult to picture a 3D model that would make this image.
202. Describe a perceptually perfect (i.e., efficient) virtual reality display. Include in your discussion. A. Visual field B. Intensity range C. Number of intensity levels D. Visual Acuity and foveated rendering E. Frame rate and update rate (what is the difference?) F. Color G. Motion and spatio-temporal blurring
<probably/definitely on exam> A. 180-190 degrees, like the human visual field B. 1:10,000 like the real world. In VR, there is still an issue of stray light, so this is difficult C. 10 bits (1000 levels) D. 100 pixels/degree - hard because the screen is so close, and it would need tiny pixels. It would also need a lot of power to render, so you should use foveated rendering to cut down on power. E. FR: hardware of how fast successive frames come up. UR: how fast the computer can render a new screen. Both should be ~240 fps F. currently trichromatic - can't display colors outside of this triangle. We need to either add more primaries or expand the distance between the current primaries more G. Fast movement is jerky without a high enough frame and update rate - add spatio-temporal blurring to smooth jerky motion, because it simulates how we don't have full resolution when we move in the real world
A computer screen has the standard three phosphors (RGB). Use the CIE diagram to show the approximate colors that would be produced by the following RGB outputs. What are the xyz coordinates? RGB 250 250 0 = ? RGB 125 125 0 = ? RGB 25 00 125 = ? RGB 100 00 100 = ?
RGB [250, 250, 0] = XYZ [0.45, 0.45] - yellow RGB [125, 125, 0] = XYZ [0.45, 0.45] - same yellow as before RGB [25, 0, 125] = XYZ [0.3, 0.15] - purplish-blue RGB[100, 0, 100] = XYZ [0.4, 0.2] - purple right between blue and red
201. Describe the RSVP method. What does it tell us about the speed of processing?
Rapid serial visual presentation method Experimental technique involves showing subjects a set of images rapidly, all in the same place, at some fast rate, e.g. 10 images/second. Subjects will usually be able to detect the presence, or absence, of a particular kind of image when asked. Experiments have shown that the maximum rate for the ability to detect common objects in images is about 10 images/second ex: Is there dog in one of the following pictures?
Benhams top
Benham's top is an image of a spinning wheel which is partially occluded and has different patterns of lines or shapes on it. If the top spins fast enough, you will start to perceive colors in the spinning colors. It is believed that different colors are associated with some temporal code, which is what causes this phenomena. This is not explained by either trichromatic color theory or opponent theory.
199. Provide 3 examples of biometrics and 2 issues regarding privacy.
Biometrics: 1. Iris scanners 2. Fingerprint scanners 3. Blood/pulse scanners on faces (facial recognition) Privacy issues: 1. Installing security devices in public places to prevent terrorism will allow the government to track everyone going to these places, not just terrorists. Is this an invasion of privacy? 2. They can create targeted advertisements on billboards based on your age and sex, which can be very uncomfortable
217. What is defensive design? Provide 2 examples
Defensive design is design used to prevent certain actions. 1. Uncomfortable benches in fast food restaurants so people don't stay long 2. Putting rails on benches to prevent people from sleeping on them 3. Putting spikes on signs and buildings so pigeons don't land there
209. What is deep learning? Is this a supervised or unsupervised learning algorithm?
Deep learning is a supervised method of machine learning. It involves designing a network with multiple hidden layers. The network takes in input (images for a CNN) and labels for these images (so it is supervised), and outputs a vector of the results. The network then gets feedback through backpropagation depending on whether the results were correct or not, and it updates its internal weights accordingly.
196. Describe the basic training required for a deep network to learn to classify objects. How does this differ from how most animals learn?
Deep networks are given data, make a prediction, and are then told if they are right or not. They then update their weights to correct their predictions. This has to be done many times with a lot of data in order for networks to learn. Animals mostly learn in an unsupervised manner (they don't have someone telling them the 'labels' of the objects they see all the time). They learn by probing and interacting with the environment. Animals also have a lot more feedback than deep neural nets.
204. What is the difference between demographics and psychographics? How did Cambridge Analytica take advantage of this?
Demographics- description of quantifiable descriptors like ethnicity, age, employment, etc Psychographics - non-quantifiable descriptors like values, personality, lifestyle, opinions of a person Cambridge Analytica used data from Facebook and other sources to learn a lot about both demographics and psychographics of users, using things like 'likes', shares, videos/ads watched, friends' 'likes', etc. There was so much data, that they could get a better prediction of a user's personality than their user's spouse and family
Use the CIE diagram to describe the range of colors seen by a dichromat. What is represented by a confusion line?
Dichromats are missing either their red, green, or blue-sensing cones. Looking at the CIE chart, which uses R, G, and B primaries, a dichromat would be entirely missing one of the primaries. They would only be able to see the range of colors between the two primaries they are still sensitive to.
213. What are the epistemic actions and how do they relate to perception? (W)
Epistemic actions are actions we take intended to help us discover information (eye movements, scrolling, clicking on a link, etc). Eye movements are the fastest epistemic action, so an effective display would have most or all of the information available in high resolution within your field of view, so you only need to move your eyes around or maybe your head.
According to Livingstone, why do equiluminant colors look unusual?
Equiluminant colors generally appear to be on the same surface (co-planar hypothesis), so they tend to blend more easily. They will appear to blend until they hit the 'border' of the shape they are in.
What is wrong with the description of some individuals as 'color blind'?
Everyone is technically 'color blind'. We only have 3 types of color receptors, while some animals have many more.
206. What is FACS theory and how does this relate to Avatars? (W)
FACS = Facial Action Coding System FACS is a method for measuring and defining groups of facial muscles that make certain expressions. The shape and motion of the eyes and eyebrows are some of the most important ones. This helps animators design emotions for computer avatars. Using simplified FACS muscle groupings, they can create avatars that show recognizable facial expressions that we are used to. Even simple emoticons can show some generic facial expressions to show emotion
173. Describe four ways of physically interacting with a virtual world. Which of these involve force feedback? What is force feedback?
Force feedback is some physical response that you get when you interact with the virtual environment. For example, if you pick up a ball in VR you should feel some pressure on your hands to simulate the way you would actually feel picking up a ball. Four ways of physically interacting with the virtual world: 1. Moving around in the virtual world should give you a sense of your physical location within it 2. Data gloves can track your hand movement to let you interact with objects in the world, and may provide force feedback when you touch things based on the force they would actually exert on your hand in real life. 3. Special socks/floor surfaces or treadmills to allow you to move around while staying in the same physical location in the real world. This kind of gives you the force feedback of the world, but we don't have anything yet that adjusts the feedback on your feet based on the surface you are walking on. (This is force feedback) 4. Flying VR example - you can push on the wings to 'flap' and tilt the platform to move around the sky. This has force feedback because it gives you more resistance as you flap. 5. Regular remote controls can let you control items in the environment but do not have force feedback
195. What is the multi-tasking illusion and how does this relate to switching costs?
People who multi-task are actually less productive than they think they are because of the switching costs associated with changing your focus onto a different task. It takes a little time for your mind to focus on something new, so switching between a lot of tasks actually wastes a lot of time.
Describe and provide examples of four pictorial cues, two ocular cues, and two motion cues, for the perception of depth.
Pictorial cues: - Familiar size: relates to how we are used to the size of certain objects. Example: you know how tall people tend to be, so you can estimate their distance. Another example: Disney castle plays with our expectations of house and brick sizes to make it look bigger. - Atmospheric Perspective: the air reflects some bluish light, so the more air there is between the viewer and the object, the more blue it will look. Example: As mountains get farther away, they look more light blue/gray. - Linear Perspective: parallel receding lines meet at a vanishing point. Example: MC Escher's paintings look realistic because he uses 2- and 3-point perspective to create his illusions. - Detail Perspective: things that are about the same size appear to be smaller and closer together as they get farther away. Example: The bricks on the Disney castles use this fact to make the illusion that the castle is bigger. Also bricks on a sidewalk appear to have less spacing between them and look smaller as they get farther away. - Interposition: the overlap of objects with a T junction. The top of the T appears to be the object that is in front. Example: MC Escher's illusions that make objects look like every part is in front of the other. Also hot air balloons one in front of the other have these T junctions. - Shape from Shading: the assumption that light comes from above helps give us some idea of shape. Example: Looking at a picture of a canyon from above with the light coming from below can make it look as though it is sticking out of the ground instead of cut into the ground. - Height in Plane: objects that are higher up on a surface appear to be farther away. Example: In a photograph of two people at different spots in a room, the person who is farther back will be 'higher up' on the plane of the floor. - Depth of Field: focal length; things that are at the focal length are in focus, and things in front/behind are out of focus. Example: If you hold your fingers one a few inches behind the other, the one in the middle of your view is more in focus and the other one becomes out of focus, no matter which one you look at. Temporal/Motion Cues: - Motion Parallax: most important depth cue (more than disparity). Objects that are closer to you seem to move faster and in the opposite direction of your motion, while objects that are far away appear to move slowly in the same direction as you. Example: If you walk across campus, street signs that you pass move by you fairly quickly, while buildings that are far away seem to move in the same direction but slowly. - Rate of Acceleration: the speed of an object depends on its weight and size with the rate D = 1/2 * a * t^2. Example: Godzilla is huge, so its limbs will move much more slowly than a person who is running away. In film, you would need to speed up the filming by a rate of sqrt(desired size / actual size), and then play it back at the opposite slower rate in order to create this effect. - Kinetic Occlusion: tells you the relative size of two objects. Example: ??? - Kinetic Depth: structure can come from the motion of an object. Example: a silhouette of an object may be indeterminate, but if it starts 'spinning' you can get a better idea of its structure. Ocular Cues: - Binocular Disparity: Each eye sees a slightly different image because they are at different angles. Example: If you hold up a book right in front of your face, it will look close because your eyes are seeing two images that are relatively different. If the same book is held 100 feet away, it will look far because each eye sees approximately the same image of the book. - Accommodation: The eye lens can control the focus of the image you are looking at to make it look in focus. The lens will distort different amounts to focus on an object that is nearby or far away. Example: If you look at your computer your eyes will focus on the screen, but if you look out the window your lens will change shape to accommodate the change in focal length. - Convergence: Our eyes converge (point slightly inwards) as we look at objects that are nearby. Example: This can help us see some optical illusions, like Magic Eye, which focuses our eyes in front of the image and allows two parts of the image to overlap in our visual system, showing the hidden image.
McAdam Ellipses
McAdam ellipses are regions on a color chart that we can discriminate equally. Since we are more sensitive to changes in blues than greens, for example, the ellipses in the blue regions of the color chart are smaller than the ellipses in green regions of the color space.
In terms of chromaticity and luminance discuss the limitations of the range of possible colors on television, film, and the real world. Why is white similar across these media?
Most TVs and screens are limited to the RGB color space, meaning that all of the colors they can produce are limited to the linear combinations of red, green, and blue. In terms of luminance, color (except for white) in displays is created by filtering out some light. This means that luminance is tied to chromaticity. The intensity of a color on a display is limited by the intensity let through that color filter. White is similar across media, because it is always the brightest color- TV: white = 256 units of light from R,G,B- Film: white lets in the most light- Real world: white reflects the most light
203. Describe two experiments with priming that provide insights into object recognition. (W)
Priming: You are more likely to recognize something if you have just seen it recently 1. Bar & Biederman - showed subjects pictures so quickly they couldn't recognize them. They then flashed a visual mask to remove the image from the visual iconic store. After 15 minutes, the chance of recognizing the original image were more than by chance. 2. Lawson, et al. - Subjects were shown a set of pictures and had to recognize a certain object. They were much faster if they had been primed with images that were visually similar, but NOT of semantically-related objects.
Explain the concept of negative primaries in a color matching task. Imaginary primaries or negative light?
Negative light: If you choose the primaries of your color space to be within the range of visible colors, then you can describe all visible colors using combinations of these primaries. Some of the possible colors will require negative amounts of your primaries because they fall outside the convex hull described by the three primary axes. Imaginary primaries: If you choose the primaries to be outside of the color space, then they are 'imaginary' primaries, but you can use combinations of positive amounts of your primaries to create any color within your color space. You will also be able to describe other colors outside your color space with all positive amounts of your primaries.
174. If you were to take a vacation using current VR software, what aspects would you be missing? What aspects are likely to change in the next 10 years?
No senses such as smell and feeling the weather (i.e. rain, heat, cold, wind). Some of these may be added in the future (i.e. the fan used to blow wind and the scent tubes in the flying VR system). You couldn't do certain activities like swimming, and currently force feedback wouldn't be able to handle activities such as playing frisbee or catch with a ball because it can't handle that type of force that pushes your whole body back. Technology is constantly improving (resolution, update rates, force feedback, etc) so those will improve in the system.
Corteen and Wood experiment
Participants were presented with different words and city names. They received an electric shock whenever a city name was presented. Then they had to do a shadowing task, where they had to repeat a passage read aloud to them in one year. In the other ear, they were given names and city names. Their galvanic skin response was measured, and found to be highest for city names and lowest for non-city names.
What is the difference between protanomolous and protanopia?
Someone who is protanomolous has abnormal red-sensing cones, while someone with protanopia is completely missing their red-sensing cones.
184. What is the evidence that subliminal perception exists? What are the conditions under which it might be possible? Under what conditions is it unlikely?
Studies have shown that flashing a stimulus to a person at a speed where they can detect it but not recognize it consciously can affect their later behaviors. Something that is not very likely to be recognized, but is very likely to have an effect on your behavior, is subliminal perception. For example, the mere exposure effect showed that being exposed to an image so briefly that you couldn't name it later can still make you like it more than an image you hadn't seen. This is not likely to work on things you already have a strong opinion are (i.e. celebrities or politicians)
187. Under what rational definition of subliminal perception can we say that it exists?
Subliminal perception is when you are subjected to stimuli that are flashed quickly enough for you to detect, but not recognize or be conscious of, and it has some effect on your behavior. Something that is not very likely to be recognized, but is much more likely to have an effect on your behavior, is subliminal perception. It has been shown that stimuli can cause some cognitive effects if shown enough to be detected but not recognized. However, - this requires precise timing - effective unattended stimuli are often easily recognized when attended - Little evidence that it is more effective than supraliminal stimuli
194. What is synesthesia? Describe one line of evidence that synesthesia is a 'real' perceptual phenomenon.
Synesthesia is a condition where people associate their senses in a way that is not actually in the stimulus. Most commonly, people may perceive numbers and letters has having certain colors (i.e. 3s are always green). Other examples are hearing colors, associating moods with different numbers, or tasting different shapes. Some evidence that synesthesia is a real perceptual phenomenon: - People with synesthesia are faster at certain search tasks. For example, if you are shown a page with a bunch of 2s and 5s in black, it would be slow to find all of the 2s. A person with synesthesia would find the 2s much faster because the 2s and 5s may be different colors to them - it is involuntary and automatic; people with synesthesia can't 'turn it off' - it is consistent; people with synesthesia always have the same perceptions (i.e. 3s are always green, it doesn't change every time they see a 3)
177. What is the TVSS? Briefly describe Guarniero's experience with it. (See readings)
TVSS is the Tactile Vision Substitution System. It was invented by Guarniero, a blind doctoral student at NYU to help him perceive the visual world through vibrational sensors on his back. After a while of practice, he described the sensation as unusual, saying it was clearly not sight but also no longer felt like vibrating pins on his back. Guarniero discussed some of the challenges being: - difficulty differentiating whether he is perceiving motion because he was moving, or the object was moving - holding the camera level or adapting to how much it was tilted by recognizing lines of different orientations At first, it took time to scan each object in order to recognize it. After a while, he was able to extract some distinguishing features in order to recognize items more quickly. After switching over to the better camera, he also had to learn the affects that each zoom level had, and eventually he regained familiar size. With more practice, he was able to distinguish multiple objects, and say which was farther away than the other, mostly using interposition and height-in-plane cues. The addition of the mirror allowed for more depth cues, such as convergence of parallel lines. With more practice, he was also able to get camera-hand coordination to interact with the world properly.
Bruneleski in 1413 described what primary aspect of linear perspective?
That parallel lines converge to a vanishing point
188. How does the concept of affordances relate to effective design? Give three examples.
The concept of affordances is the design of something can suggest certain actions which have some natural mappings to human behavior. For example, a chair affords sitting, a pencil affords writing. Something that has good design should afford the actions that it intends to have. 1) A flat bar on a door affords pushing because it an easy object for us to push. If it had a pull bar, it would not afford the action necessary to use it, so it has bad design. 2) A cap on a bottle affords twisting because we can hold the bottle in one hand and the top in the other and our fingers can spin to turn the top off 3) An X on a computer window/tab affords clicking to close 4) A mug handle affords grabbing and picking up
189. Recognition of a face appears to have a number of stages. Explain. Provide one line of evidence for these stages.
The four stages are: 1. Recognizing the face is a face 2. Recognizing the face is familiar 3. Retrieving stored biographical information 4. Retrieving the person's name In experiments, people have shown that answering questions regarding later stages takes longer than answering questions in previous stages.For example, people will often be able to see a celebrity face and know what they do (actor, musician, politician, etc) but not remember their name. It rarely happens the other way around, where you know the person's name but not their occupation or some other facts about them. This suggests that it takes longer for us to process information such as a person's name than it does for us to process that we are seeing a face.
In the video showing two lamps playing with a ball, why do the lamps look alive and seem to have personality?
The lamps look alive because they have biological motion. This is characterized by having smooth, not jerky, movements. This includes slowing down as object direction changes (i.e. move slowly at high-angle curves). For example, rather than falling flat on the ground when the lamp jumps, it bends in a smooth, continuous motion in order to break its fall, just as a human would.
197. Describe an experiment showing the sensory identification is limited to 7 plus or minus 2 levels.
The magic number, 7+/-2, refers to the number of stimuli we can process at a time. In one example, you can show people different shades of gray boxes, hide the boxes, and then show them only one of the previous boxes they found. Most people are successful at saying which shade of gray from the previous set of boxes was shown when there were originally between 5 and 9 boxes.
According to opponent color theory what causes the perception of white?
The perception of luminance is described as one of the opponents (white vs black). A perceived white is created by balancing the opponent colors (R-G and Y-B)
What is the disappearing hand trick, and how does this relate to visual capture?
The person puts their hands into a box that has cameras inside. There is a display on the top of the box that shows images of the two hands in the box. The display shows the hands getting closer together, even though they are not. The user then drifts their hands apart to keep them from touching. The hands keep getting farther apart, while the video just shows the hands about the same distance, getting closer together every now and then to encourage the user to move them apart more. Visual capture is the phenomena that our visual input is dominant over our other sensory inputs. In the disappearing hand trick, the people really think their hands are side by side, because that is what is shown in the video, even though they could feel that their hands are far apart.
The cocktail party effect
The phenomenon that occurs when you hear your name being said from across a crowded, noisy room. It suggests that we still hear what everyone is saying, but we only attend to some of it, when it is important to us.
198. You have been asked to design a display to monitor racecars on a race track. The display should provide an aerial view of the track and allow the observer to quickly identify each of the 8 racecars, provide a cue as to the amount of fuel left and warn the observer when the fuel in any of the cars is critically low. Discuss the perceptual problems involved in designing such a display and a possible solution.
This task is difficult because it requires showing a conjunction of features, which makes information slower to find. At any one time, the user has to track at least 16 things (the 8 cars and the 8 corresponding fuel levels, as well as if any of the cars are low on fuel). You want to make the search faster and parallel by separating different things by a single feature rather than a conjunction of features. To design such a system, I would recommend using color coding to match the cars (maybe a colored bounding box) to their fuel gauges of the same colors. This way, matching a car to its fuel level only requires single feature processing rather than conjunctive feature processing. I would then make another simple, single feature to indicate whether a car has low fuel. Perhaps drawing an exclamation point over the car or the fuel gauge. This way, users only need to check if there is or is not an exclamation point by the car, which is simple and can be done in parallel (single feature).
181. How does one generate a caricature? How is this relevant to prototypes? How might this relate to the difficulty in identifying a face of a different race?
To create a caricature, you would exaggerate some distinguishing features of a person. For example, if a famous person wears big glasses, you might give them huge classes in a caricature. Because we recognize people based on their variation from the facial prototypes we have in our memories, exaggerating distinctive features can actually help us recognize people more quickly. In fact, people will sometimes think a caricature looks more like the person than a true depiction of the person. When we create prototypes of faces, they are based on the faces we have been exposed to in our lives. A prototype is the average of all of the facial features we have seen. If we have had little exposure to people of a certain race, this may create a poor prototype of that race in our minds, making it harder for us to recognize faces of that race than ones we have had more exposure to. After you are exposed to more faces of that race, you can create a prototype of that race and be better at differentiating them.
186. Describe the difference between virtual reality, augmented reality, and mixed reality. Provide an example of each.
VR: Creates complete world with HMD system. No visual input from the real world. (HTC Vive, Oculus Rift) AR: Provides a view of the real world with some computer generated components not typically anchored to the real world. (Google Glass) MR: Provides a view of the real world with objects that are anchored to the real world (Pokemon Go)
190. What is negative priming?
When you focus on something, you are inhibiting other nearby stimuli, slowing down your ability to focus on the stimuli you were inhibiting before. Example: overlapping letters, reading the red ones over the green ones, but each red letter is the same as the last green one.
Metamer
two things which look the same, but are physically different. For example, a pure color purple can look the same as a purple produced on a display which uses red and blue pixels.
