Chapter 6 Chapter 4 Visual Psychology

"When we trace the history of the development of psychology, we run into a labyrinth of fantasy, contradiction, and fallacy intertwined with certain truths." —Thomas Reid I hope I've convinced you that seeing is not as easy as you might think.Seeing is a constructive process in which the brain responds in parallel to many different "features" of a scene and combines them into a meaningful whole, guided by past experience.Seeing involves certain active processes in the brain that lead to a clear, multi-layered, symbolic interpretation of the scene. We will now consider some of the basic operations that the brain must perform when we look at an object, its position relative to us and other objects, and a certain attribute of its shape, color, motion, etc.Perhaps the most important thing we should realize is that objects in your field of view are not what you see them to be.Each object is not marked in a clear and definite way, and your brain must use various cues to bring together the parts of the scene that correspond to the same object.In the real world, this is not an easy task.Objects may be partially occluded or appear against a confusing background.

An example will make it clearer.Take a look at this photo in Figure 5.You immediately see, effortlessly, that it is the face of a young woman looking out the window.But if you look closely, you will find that the wooden window lattice of the window divides the woman's face into four parts.But, you don't see it as four separate pieces of four different faces.Your brain puts them together and interprets them as a single object—a face partially obscured by the wooden sash in front of you.How did this combination come about? This is Gestalt psychologist Max Wertheimer.One of the main research interests of Wolfgang Kohler and Kurt Koffka.The movement began in Germany around 1912 and ended in the United States.All three of them left Germany when the Nazis came to power.My dictionary defines "gestalt" as "an organic whole whose parts interact so that the whole is greater than the sum of its parts." In other words, your brain has to draw on your past experience and your The experience of ancient ancestors embodied in the genes of human beings actively constructs these "wholes" by discovering the optimal combination of various parts.This combination most likely corresponds to relevant aspects of an object in the real world.Clearly, what matters is the interaction of the parts.The Gestalts attempted to classify the types of interactions common to the visual system and termed them the laws of perception.Their laws of composition include proximity, similarity, good continuity, and closure.Let us discuss them in turn below.

The law of proximity states that we tend to group things together that are close to each other and farther away from other similar objects.This is evident in Figure 6.The figure consists of many regular rectangular arrays of small black dots.Your brain may organize them into horizontal or vertical lines.But actually, you see them as vertical lines.This is because the distance from a point to its closest point is shorter in the vertical direction than in the horizontal direction.Other experiments have shown that the proximity law generally refers to "spatial proximity" rather than retinal proximity.

Gestalt's law of similarity says that we group things together that have obvious common properties (such as color, movement, direction, etc.).If you see a cat running, you put its parts together.Because generally speaking, when a cat runs, its parts move in one direction.For the same reason, a cat crawling through a bush would be identified.But if it's motionless, it's hard to spot it. A good continuity law can be illustrated by Figure 7.The upper part of the graph shows two intersecting curves.We do see it as two lines, not four lines meeting at one point or two close Vs as shown in the lower part of the figure.We also tend to think of interrupted line segments as continuous lines partially obscured by an object.

Consider the group of eight odd-shaped objects shown in Figure 8a.The middle two resemble the letter Y, and the other six are twisted arrows.Whereas in Figure 8b, you will probably see a 3D cube frame obscured by three diagonal bars.Now, those grotesque objects have become an integral part of the upper and lower pictures.The cube is easier to see in the second figure because it appears to be a single object obscured by the diagonal bars.While the first figure, lacking any occlusion clues, is more likely to be seen as eight separate objects. Closedness is most evident in line drawing graphics.If a line forms a closed or almost closed figure, then we tend to think of it as a figure surface surrounded by a line rather than just a line. ①

The Gestalt school also has a general principle called "succinctness" (Pragnanz), which can be approximately translated as "goodness".Its basic idea is that the visual system makes the simplest, most regular and symmetrical interpretation of the input visual information.How does the brain decide which explanation is "simplest"?The modern view is that the best explanations tend to require very little information (in the technical sense) to describe, while the bad explanations tend to require more information. ① In other words, the brain needs a plausible explanation rather than a fancy one.This means that the interpretation does not change fundamentally with small changes in observation point.This is because, in the past, when you looked at an object, you used to move through the scene, so your brain had registered various aspects of the object and believed that they belonged to the same thing.

The laws of Gestalt perception cannot be regarded as strict laws, but can only be regarded as a practical heuristic research.Therefore, they can serve as a suitable gatekeeper for vision problems.What operational processes actually lead to the emergence of these "laws" is what many visual psychologists have tried to discover. As the Gestalt schools have recognized, an important operation in vision is figure background separation.The object to be recognized is called the "figure" and its surroundings are called the "background".This separation may not always be easy, and a careful look at Figure 9 will show that if you have never seen this image before, you will have a hard time seeing any recognizable objects.But after a while, you might realize that part of the drawing represents a Dalmatian dog.In this case, the separation of the graphic background is intentionally complicated.

It is also possible to construct an ambiguous graphic background separation image.Please see Figure 10.At first glance, it looks like a vase, but on closer inspection it might appear to be a side view of two faces.Originally the vase was a figure, but now the outline of the face has become a figure, and the original vase has become the background.However, it is difficult to see both explanations at the same moment. When the brain decides which visual features belong to an object, it relies on salient visual cues that roughly follow the Gestalt laws of perception.Thus, if an object is solid (proximity), has a definite outline (closedness), moves in one direction (common fate), and is red throughout (similarity), then we are likely to perceive it as a Sports red ball.

It is vital for an animal to excel at such tasks.Otherwise, it will have a hard time spotting predators or prey and other foods like apples.It must be able to separate the graphics from the background.So-called camouflage is an attempt to confuse this process, the effect of camouflage is to break the continuity of the surface (such as the camouflage uniform worn by soldiers) and produce an easily confused silhouette, so that the real silhouette is camouflaged.Colors may also blend with the background.A tiptoeing cat pauses now and then to avoid giving its prey any clues of movement.As has been suggested, our evolved good color vision enabled our primate ancestors to spot red fruit against a background of turbulent green.The thing that gives us so much visual pleasure may be the device that first finds food and sees through camouflage.

What we know about the earliest stages of visual processing comes in part from studies of the eye and brain (see Chapter 10).Pretty much the earliest operation that needs to be performed is to remove redundant information.Photoreceptors in the eye respond to the intensity of light falling on the eye.If you looked at a perfectly uniform and smooth white wall, all the photoreceptors in your eyes would respond to the light in the same way.Is there any reason to send all this information to the brain?For the fundus retina, it is best to process this information first, so that the brain knows where the light intensity changes in space—the edge of the wall.If there is no change in light intensity across the entire retinal area, then no signal is sent.From "no signal" the brain can infer "no change" and that this part of the wall is uniform.

As we will see in later chapters, the brain's processing of different types of information is, to some extent, carried out in parallel pathways.So it makes sense to study separately the processes of how to see shape, motion, color, etc., even though these processes interact to some degree. Let's start with the shape first, obviously, extracting contours is very useful for the brain.This is why we respond so easily to line drawings.Even without any shading, texture, color, etc., you can still interpret the line graphics of a scene (Figure 11).This suggests that some elements of the brain respond better to fine details, others to less detailed parts, and others to coarse changes in space.If you could only see the latter, the world would be as blurry as out of focus.Psychologists often use the term "spatial frequency".High spatial frequencies correspond to fine details and low spatial frequencies respond to slow spatial changes in the image. Please see Figure 12.You'd probably see it as a collection of small squares with a uniform grayscale.Now, if you blur it (take off your glasses, close your eyes half-closed, or place it far away in the room), you might recognize Lincoln's face.The details of the figure (the edges of the small squares) interfere with the recognition process.These details are less noticeable when vision is blurred.At this time, although the image is still somewhat blurred because there is only low spatial frequency information in the image, you can recognize his face. helpful. One of the most difficult problems the brain faces is extracting depth information from two-dimensional images.We need depth information, not only to determine the distance of objects from the viewer, but also to recognize the 3D shape of each object, and using both eyes is helpful.But its shape can often be seen with just one eye or by looking at a picture of it.What cues does the brain use to derive three-dimensional information from two-dimensional images?One clue is the shadowing of objects created by the angle of incoming light.Please see Figure 13.You might see one row as four depressions in the plane, and the other row as four protrusions.This impression of depth comes from the shadows of people shooting light. Occasionally, this interpretation can also be ambiguous.Stare at the diagram for a moment, or turn the page upside down, and you'll see the indentations as protrusions, or the protrusions as indentations (note that this happens at the same time).Your brain initially thinks that the lighting is coming from one side, but if the lighting is actually coming from the other side, then the same shadow will correspond to a different shape, as you can see. Another convincing lead is "recovering structure from motion".This means that if the shape of a stationary object is difficult to see (often due to the lack of some three-dimensional shape cues), turning the object slightly makes it easier to identify.It is difficult to understand when a model of a complex molecule made of spheres and spokes is projected on a screen during a lecture.But if you play a movie of its rotating model, its three-dimensional shape becomes clear at a glance.You may have seen this scene at the end credits of the TV show "Story of Life."There, models of DNA molecules spun to music in the air. To perform three-dimensional observation, it is not enough to see every object in three-dimensional space.You also have to watch the entire scene in three dimensions in order to figure out which objects are close to you and which ones are far away.Even 2D images have two strong depth cues. The first clue is perspective, which can be vividly demonstrated with the Ames Transformation Room (named after its inventor, Adelbert Ames).Such a room can only be viewed with one eye from the outside through the small hole.In this way, any stereoscopic cues can be ruled out.The room looks like a cuboid, but in reality it is long on one side.One of its corners is much higher and farther away than the square room.When I looked at such a room through a small aperture at the Exploratorium in San Francisco, I saw some children running around in the room.On one side of the house they appear tall (when they are close to me) and on the other side they appear short (when they are far away).As they run from side to side (actually from a close corner to a far corner and back again), their size changes dramatically.Of course I understand that it is impossible for children to change their height in this way.But the illusion was so real that I couldn't shake it off right away.The apparent size of each child is created by the false perspective of the walls.Like other illusions, this one is difficult to correct with "top-down" (ie, the highest levels of the brain understanding that forms the basis of this illusion) effects. Another strong clue is occlusion.i.e. an object close to you partially occludes a distant object.We have already seen this scenario in Figure 5.A girl's face lies behind the frame of the window pane.Using this clue, the brain can infer that the various parts of the occluded object should belong to the same object, as we discussed at the beginning of this chapter. Lines can produce two magical effects related to occlusion.The Kanisha triangle shown in Figure 2 belongs to the first type.The phantom boundary of the white triangle is formed by the extension of the straight boundary of the black defected disk.Another effect is shown in Figure 15. The phantom boundary in this case is mainly due to the fact that the endpoints of the group of line segments line up."Lines" in airports can occur for a variety of reasons, such as the pattern of an object (such as a shirt) or the stripes of a zebra and shadows.An object that occludes the background often cuts off lines in the background.In this case, the phantom contours created by the endpoints of the line segment will outline the object, as in the deliberately contrived figure of Figure 15.As the psychologist VS. Ramachandran put it: "The perception of phantom contours may be more real (more important to us) than real contours." Another distance cue is the gradient of the texture.As shown in Figure 16.As long as you see this picture of grassland, you will immediately have the impression that the grassland is gradually moving away from you.This is because the blades of grass on the page gradually become smaller from bottom to top.Your brain doesn't see it as a flat, vertical wall below which the grass grows taller and above which grows lower: instead, it sees it as a wall of uniform height stretching out into the distance. lawn. There are also some depth cues.One is the apparent size of the object.A familiar object has a smaller retinal image as it moves away from us.Therefore, if the apparent size of the object is small, the brain perceives it to be farther away.Another depth cue is that distant landscapes often appear bluer.All these clues were exploited by artists, especially after the discovery of perspective during the Renaissance.Canaletto's Venetian landscapes are a good example. Let us turn to discuss the main sources of depth information (1).It is often referred to as "stereoscopic" and relies on small differences in the scene image when two eyes view the same object. Physicist Sir Charles Wheatstone was the first to clearly demonstrate in the mid-19th century that a properly rendered binocular image can give a vivid impression of depth. (There is another interesting incident about Whitestone that stands out. He once ran away due to high tension while waiting to deliver a Friday night speech at the Royal Society in London. Since then, every speaker has been locked up before the speech as a rule. Wait a quarter of an hour in a tiny house.) Whitestone also invented the stereoscope (popularized after the war for its simplicity of design).It makes it possible for each eye to look separately at a photo taken from a slightly different angle.Differences in shooting locations produce landscapes that are not strictly identical.The brain detects the difference between the two landscapes (this is technically called "parallax"), and the result is a scene in the photo that appears to have a distinct sense of depth, seemingly right in front of you. When you observe the real scene that is nearer in front of you, you can experience what stereoscopic vision is by closing one eye.For most people, the perception of depth is not as strong as when using both eyes at the same time. (Of course, you can still have a good sense of depth even with one eye closed due to the other depth cues mentioned above.) Another obvious example is sketching or photography of architecture, cities, landscapes, etc.In this case, two eyes allow the brain to deduce that the picture is flat.In fact, a more vivid depth perception can still be obtained with one eye.As long as you stand in a position where there is no reflection from the glass, and block the frame of the picture with your hand.These actions remove some planar clues on the surface of the picture, making the clues that the artist uses to express depth information in the picture produce a stronger effect. Stereoscopic vision is most pronounced for objects that are closer to you because binocular disparity is at its greatest.Obviously, for both eyes to see the same object, the object has to be almost directly in front of you.It should not deviate so far to one side that the nose obscures the view of one eye.Animals that rely on predation, such as cats and dogs, usually have both eyes facing forward.This allows them to use stereovision to catch prey.For other animals, such as rabbits, it is more beneficial to have eyes on both sides of the head, so that they can spot predators in a wide field of vision.But their stereo vision is limited compared to humans because the fields of their eyes overlap very little. ① What about sports?The reasons why the visual system is interested in motion are obvious.When you watch a movie, although what you see on the screen is a series of rapidly rendered still images, you have a vivid impression of moving objects.This phenomenon is called "apparent motion".In this rather artificial situation, the visual system may err.The spokes of a car or wagon wheel sometimes appear to turn in opposite directions.Generally speaking, the reasons for its occurrence are already clear.This is largely caused by the brain associating one spoke in one image with the one closest to it in the next.Since the wheel is constantly turning, it may not be the same spoke that is linked together, but other adjacent ones.Since all the spokes look exactly the same, the brain is likely to associate two different spokes in two adjacent images together.If the two spokes connected together were in exactly the same position (relative to the car), the wheel would appear to be stationary.If the revs are slowed down even slightly, the spokes of the wheel will appear to turn backwards.Especially in old movies, this phenomenon happens from time to time.When the car slows down, the spokes appear to change direction (relative to the car's motion).Psychologists have performed numerous experiments in an attempt to determine what is required to achieve good apparent movement. Another motion effect is the barbers pole illusion.Because the cylinder has helical stripes, when it rotates around its long axis, the stripes don't appear to be turning but to move along its long axis, usually upwards. (This will be discussed fully in Chapter 11. Therefore, our perception of motion is not always direct. In this case, what you see is not a localized movement of each stripe, but a brain error. Think of it as a global motion of the entire pattern. Motion perception in the brain is processed by two main processes.They can be loosely referred to as "short-range systems" and "long-range systems".The former occurs at an earlier processing stage than the latter.Short-range systems do not recognize objects, but only changes in light patterns received by the retina and transmitted to the brain.It can extract the "primitives" of motion, but it doesn't know what objects are moving.In other words, this simple motor information is useful as a primary sense.It operates automatically, i.e. it is not affected by attention. It has been hypothesized that short-range locomotion can use motion information to separate patterns from backgrounds① and is related to the after-exercise effect (sometimes called the "waterfall effect"). (If you stare at a waterfall for a while, and then quickly move your gaze to an adjacent rock, for a short period of time, you will see the rock moving upward.) There are different views on this phenomenon.Because recent experiments have shown that post-exercise effects can be influenced by attention. The long-range locomotion system appears to be involved in the register of object motion.It not only registers the movement itself, but also what objects are moving from one place to another.The long-range motor system is affected by attention. Let's take an (oversimplified) example.A red square flashes on the screen for a short time, and after a while, a flashing blue triangle appears not far from the red square.If the time, distance, etc. parameters are chosen such that the long-range system dominates, then the observer will see the apparent motion of the red square changing into a blue triangle and moving from one location to another.On the other hand, if the parameters are chosen to excite mainly short-range systems (small time intervals and distances), then the observer will only see motion and not the moving object.He feels movement but doesn't know what is moving.In most cases, both systems may work to some degree.Only well-designed stimuli activate just one system. The brain uses motion cues to gain additional information about the changing visual environment.I have described how in some cases structure can be recovered from motion, but motion information can also be exploited in other ways.An object that is running toward your eye creates a progressively dilated retinal image.If an object on the screen suddenly grows larger, you will feel that object is rushing towards you (even though the screen is still at the same distance).This visual image movement is called "dilation".The effect it produces is so vivid that one suspects that a special part of the brain responds to the expansion of the image.In fact this site has already been discovered (see Chapter 11). Another role of the visuomotor system is to guide the way you move around your environment.When you walk forward, your eyes look ahead, and your visual scene of up, down, left, and right will pass you by.The movement of this retinal image is called "visual flow", and it is of great help to the pilot when the plane is landing. A monocular pilot without stereo vision can use the visual flow information to land the plane safely.The place where there is no visual flow is the point you are moving towards it.All objects around that point appear to be moving away from it, although their velocities vary (see Figure 17).This visual information helps pilots find the correct landing spot on the runway. Color perception is also not as straightforward as it seems.The basic idea is that it has to do with different types of photoreceptors in the eye.Each photoreceptor only responds to light within a limited range of wavelengths.It is important to realize how the response of individual photoreceptors does not depend on the wavelength of the incoming photon.A photoreceptor may or may not capture a photon.If it does get caught, the effect will be exactly the same regardless of the photon's wavelength.But the probability of its response depends on the wavelength.Certain wavelengths have a high probability of activating it, and some wavelengths have a low probability of activating it.For example, it can often respond to "red" photons but rarely respond to "green" photons. The average response to an input photon stream may correspond to a few photons in a sensitive band, or to many photons in an insensitive band; the receptor cannot distinguish them.This all seems pretty complicated when you first read it, but experience tells us that if your eyes have only one type of photoreceptor, your brain loses information about the wavelength of light and can only see the world in black and white.This occurs during extremely dark times, when a type of photoreceptor called a "cone" is inactive and only the "rod" receptors are active.These are all one type of photoreceptor, responding equally to all wavelengths.This is why you cannot see the color of the flowers in the garden when the night is very dark. To obtain color information, more than one photoreceptor with different wavelength response curves is required.Their response curves are partially overlapping.However, a stream of photons with the same wavelength excites different photoreceptors to different degrees.The brain uses these ratios of different excitations to determine the "color" of light falling on a point on the retina. As you know, most people have three types of cones (roughly shortwave, mediumwave, and longwave cones. They are often called blue, green, and red cones).But there are also a small number of people who lack the "red" cone cells, thus resulting in partial color blindness. ①They may encounter difficulty in distinguishing red and green traffic signals. This is the basic explanation for why we can see colors.But it still needs some fixes.Here, I just want to mention the so-called Land effect (named after Edwin Land, the inventor of the polarizer).Land showed us in dramatic fashion that the color of a patch in the field of view does not depend only on the wavelength of light entering the eye from that patch, but also on the wavelength of light entering the eye from other parts of the field of view. Why is this so?The information that enters the eye depends not only on the reflective properties (color) of the surface, but also on the wavelength of the light that falls on that surface.Therefore, the women's colorful attire would be very different in the sun than in candlelight.The brain is therefore not primarily interested in the combination of reflectance and illuminating light, but in the color properties of an object's surface.The brain tries to extract this information by comparing the eye's responses to several different areas of the visual field.To do this, the brain uses the constraint (assumption) that at a certain moment, the color of the illumination light is the same everywhere in the scene.Although on other occasions they may be distinctly different, if the lighting is pink it makes everything pink to varying degrees.So the brain tries to correct it.This is why red fibers in sunlight still look red under artificial lighting.But, as we know, it doesn't look quite the same, because the correction mechanism doesn't work perfectly. Let's briefly mention some other visual constancy.An object always looks roughly the same, even if we are not looking directly at it, making it fall on a different part of the retina, and if we look at an object at different distances, the retinal image of the object may be larger or smaller Or produce a certain rotation.However, we also see it as the same object.We take these constants for granted.But a simple visual machine cannot do this unless it has the innate machinery for the task that a fully developed brain has.Exactly how the brain accomplishes these tasks is still not well understood. Motion and color have a strange interrelationship.The short-range motor system of the brain is somewhat color-blind, and it primarily sees images in black and white.This is easily illustrated with a demo.Project a motion pattern consisting of only two colors of uniform brightness, such as red and green, onto the screen.The relative brightness of the two colors is then adjusted so that they appear to have the same brightness to the observer.This process has to be done individually for each person, because your color balance point will not be exactly the same as mine. ① This balance condition is called "equal brightness". Now, if you watch a red moving object on a green background on the screen, and the two colors are adjusted to equal brightness, it will appear to be moving much slower than it actually is, and may even stop moving (especially when you look at the screen. This is especially the case when on one side).This is because the black-and-white system in your brain sees the screen as a uniform gray (since the two colors are of equal brightness), so the short-range motor system gets little motion information. All of these examples illustrate that the brain can extract useful visual information from many different aspects of a visual scene.So, how does the brain deal with incomplete information provided by the outside world?The blind spot of the eye is a good example.As we said in Chapter 3, you have a blind spot in each eye that your brain "fills in."So even if you close one eye, you can't see a hole in the blind spot in your field of view.Philosopher Dan Dennett does not believe in the existence of a filling process.In his book (Consciousness Explained), he rightly argues that "the absence of information is not the same as the absence of information." He also said, "For you to see holes, somewhere in the brain has to response to contrast: either the contrast between the inner and outer edges (but in this position, your brain is not equipped for that task), or the contrast between the front and back." Thus, he argued that there was no filling, only that there There is information about holes. However, this argument is insufficient.Because he failed to prove, the information in the blind spot cannot be deduced.He is simply stating that the brain may not be making this inference.It would also be incorrect to say that the brain definitely does not have the necessary mechanisms to do this.Careful studies of the brain have shown that it is indeed possible to have certain nerve cells for this task (see Chapter 11). Rama Tsanjung, a visual psychologist at the Department of Psychology at the University of California, San Diego, conducted an ingenious experiment to counter Dennett (everyone loves to prove philosophers wrong) by presenting subjects with a fried Donut-like yellow circular pattern (see Figure 18b).The subjects had to keep their eyes still and observe with one eye.Rama placed the yellow ring within the subject's field of vision so that its outer edge fell outside the blind spot (open eye) and its inner edge fell within the blind spot (Fig. 18b).At this point the subject reported that instead of seeing a yellow ring, he saw a completely uniform yellow disc (Fig. 18c).His brain filled in the dead zone, turning a thick ring into a uniform disc. To emphasize this result, Rama-Zanzhun placed several other similar rings into the subjects' field of vision after these images were presented (one ring surrounding the blind spot, the others elsewhere).The subject reported that he not only saw the full disk in the blind spot area, but also saw the disk "pop out" immediately.This showed that the subject's attention was immediately drawn to the disc, exactly as it would be if you opened your eyes and saw a solid disc in a random array of yellow rings.Discs that are visibly different from rings will immediately jump out in front of you.As Rama said, you do fill in the blind spots instead of just ignoring what's there.Because, how can something that is overlooked really jump out? What is seen in the blind spot is not easy to study because it is 15 degrees off the center of the gaze.As I said before, we can't see things there very clearly.Rama Tsanjun and British psychologist Richard Gregory have done an experiment called "artificial blind spots."This blind spot is closer to the center of gaze. (Dennett had mentioned this work in a footnote, but was less than happy with their results.) Even more remarkably, Ramazan and his collaborators tested it on a single patient.His problem was not with the eyes, but with a small damage to the visual area of ​​the brain.Such a patient cannot see exactly what is in the corresponding position in the field of view.This area is a blind spot.But it goes without saying that his brain fills it with reasonable guesswork from his surroundings whenever he has time to spare. The results of their experiments are illustrated in Figure 19.On a cathode ray screen there are two vertical line segments lying on the same line.一条在盲斑之上，一条在下。几秒钟后，病人就会看到一条直线完全跨过间隙。一个病人还报告说，当屏幕上的线条去掉后，他"在线的填充部分看见一个非常生动的幻象"，其持续时间有好几秒，更令人惊奇的是，如果呈现给两个病人的是两条错开的竖直线（图19c所示），开始，他们看到的是两条错开的直线，但后来两条线就会相互"漂移"靠近，最后两条直线完全对齐。然后，大脑填充上它们的间隙，形成一条连续的直线（如图19d）。报告称，这些线的水平移动（记住，它们实际上是完全静止不动的）栩栩如生。两位病人对此现象深感惊讶，并表现出极大的兴趣。 其他的一些实验表明，并非视觉每个方面的填充都是同时进行的。形状、运动、纹理和颜色的填充可以在不同时间内完成。例冤五章注意和记忆如，当视场由许多运动的随机红点组成时，一个病人将颜色"渗入"到盲区几乎是立刻完成的，而在5秒钟以后才会形成运动圆点的动态模式。 需要注意的是，大脑中因伤害形成的盲斑与眼睛真正的盲斑两者所引起的结果具有重要的区别，对于后者，填充差不多是立刻完成的。在大脑损伤的情况，这个过程则需要若干秒。这大概是由于损伤失去了大脑中快速填充的部件。 填充可能并非是盲点所特有的过程。更可能的情况是，它以某种形式发生在正常大脑的多种水平。它使大脑能从仅有的部分信息中猜测出完整的图画。这是一种非常有用的能力。 现在，我们对视觉心理学的复杂性已有了大体的了解，显然，观看并非是一件简单的事情。这与我们仅凭日常经验作出的猜测有很大的差别。它的工作方式还没有被我们完全理解。它涉及许多我们不得不略去的实验和概念。下一章我们将涉足看的两个其他方面——注意和短时记忆，用来拓宽我们的研究领域。它们都与视觉意识有紧密的联系，而且还会引人不同视觉加工所需时间这样一个十分棘手的课题。 ①正如我在第一章所解释过的，如果过于简单地理解"和"这个词，这当然是正确的。 ①最近，加利福尼亚大学（伯克利）心理学家斯蒂芬·帕尔莫（Stephen Palmer）提出另外两条律：共同区域（commonregion）和联结性（connectedness）。共同区域（或称包容性）意味着相同的知觉区域组合在一起。联结性是指视觉系统把均匀的、联结在一起的区域知觉为单一单元的强烈倾向。 ①这可能或多或少地依赖于估计信息内容时采用的是哪些"基元"（primitives）。 ①大脑如何利用视差是个值得重视的理论问题。比如，需要弄清楚，一只眼睛的图像中的哪个特征与另外一只眼睛的哪个特征相对应。这称为"对应问题"。最初认为，要解决这个问题，大脑首先要识别物体，在贝尔实验室工作的匈牙利心理学家贝拉·朱尔兹（Bela Julesz），用随机点立体图进行的精彩的实验清楚地显示，两图之间的"对应"可以在先于物体识别的、低水平的信息处理阶段实现。 ①一小部分人似乎缺少真正的立体视觉。 ①这种从背景分离图形的任务提出了一个困难的理论问题，因为大脑必须在不知道什么是图形的情况下进行图形背景分离。 ①严格他讲，我们大家都是色盲。因为除了像紫外线这一类我们不能看见的波长外。可以构造出任何数目的、在我们看来是完全相同的波长分布；而它们如果用一个合适的物理仪器去测量，实际上并不完全相同。除了少数情况有保留外，我们对任一波长分布的响应可以与仅仅三种波长的合适组合相匹配。这是早在19世纪就已确认的事实。按数学术语，颜色是三维的。 ①即使对于同一观察者，位于注视线上的物体与位于视场外围的物体，它们的平衡点也可能稍有不同。