Perceiving Shape from Shading - abstract


Ramachandran, Vilayanur S. "Perceiving Shape from Shading". Scientific American. August 1988. pp. 76-83.

Our world is a three-dimensional world, yet our vision is perceived on a two-dimensional level. How does the brain take into account features such as shading to achieve the 3-d picture?

One aspect of vision which was analyzed was the viewing of element as either concave or convex. On paper, shaded figures are presented. However, the brain is unable to determine whether they are concave or convex, because it is not sure where the light source originated. Thus, occasionally with or without a conscious effort, the viewer can alter the concavity of the image. However, when the viewer does this, all of the images change concavity.

In a similar experiment, different things were shaded differently. Some were shaded in the mirror image of the other. In this situation, when one group appeared to be concave, the other appeared to be convex and vice versa. From this experiment and the previous one, it was concluded that shading is a global, not local, function. This is primarily based upon the fact that the brain only perceives one light source (the sun). Furthermore the visual system assumes that the light is coming from above.

The brain requires more than shading to perceive its three- dimensional view of the world. It also requires a figure to be outlined. One experiment was done in which two different figures were created with identical shading patterns. The only difference between these two figures was the outline surrounding them. In this experiment, the perception of depth in the two objects was greatly different. The effect of shading on one produced a completely different effect on the other. Furthermore, the brain will often create illusionary outlines if the outlines are not supplied. In one experiment, four grey disks with 'bites' taken out of them were placed opposite each other. This picture created the illusion and outline of a sphere, even though it contains only a small number of things which actually identify the circle.

From these experiments it was concluded that the brain uses a combination of shading and outline to produce its perception of depth. However, the brain needs to do many more things to provide us with the view of the world which we perceive. The brain must be able to distinguish among the many different images that it receives. Somehow it manages to do this job marvelously. The brain is able to associate the proper shading pattern with its image, even in highly complex picture. It is also able to mentally isolate individual objects from the viewing range, and focus especially on it. How does the brain do this? In the earliest stage of perception, the brain isolates certain elementary features such as color, oriented edges, and direction movement. Once these elementary features have been found, they are grouped together to form the object that we see.

Another feature that has been associated with visual perception is the ability to detect symmetry. The brain constantly notices symmetry in everyday objects, such as faces. This perception of symmetry is based on the perception of depth, not the actual distribution of light and dark. One of the drawings included with the article portrayed two separate units. One depicted was truly symmetrical. The other was not truly symmetrical, but contained units which were perceived to have a similar depth. In the sample, the truly symmetrical drawing appears to be asymmetrical. This drawing helps to provide direct testing means for the reader of the article. [egad! The horrendous point-of-view shift!]

Shading is one of the most important aspects to visual recognition, and studies have shown that there may exist single neural channels specially designed for the carrying of shading signals. Scientists have also created a three layer simulated cell network which has been used to help discover the activities of the vision. This system was modeled after the receptive fields of a cat's eye. From this simulation, they found a direct response to the viewing of certain simple figures. However, the full relevance of this experiment cannot be realized because it deliberately excluded such important features as outlines.

The drawings included in this article were a great asset. Though this article could have probably been effective without any visuals, the pictures greatly increase its value. they allow the reader to perform a self-test, and verify for himself what is said in the article. They prove that the researches didn't get a bunch of 'freaks' for their experiments. They also provide some interesting activities for the reader, even if he is not thoroughly interested in the content of the article.

Though the drawings were exceptional, the article had much to be desired. It lacked an overall structure, and often jumped from point to point incoherently. It also often repeated itself. A careful editing would greatly increase the value of the written part of this article.

In the opening, the author states that there is much more to be done before we understand visual processing. His research barely scratched the surface, focusing on the shadings importance. There is still much more to be discovered in the area of shading, and also in the many other areas associated with vision.