Two of the theories that we have been discussing regarding object recognition are the gestalt theory and geometrical ions, called geons. I have heard of both of these theories before, and both have always struck me as overly simplistic. We are able to identify by sight countless objects that we encounter in our everyday lives. Even in the absence of color and markings that might give us additional clues to an objects identity, we are able to quickly identify so many objects, of such varying shapes, that I find it hard to believe that either of these views accurately describes the incredibly complex processes that our visual brain must go through in order to properly identify objects.
The geon theory is the less plausible of the two for me. The world we see is made up of so many objects, of such varying shape and complexity that I find it difficult to believe that it can really be broken down into a number of goens, that can be distorted to form any object that we see. Obviously the world is made up of different shapes, and certain objects like a phone, mug or briefcase like we looked at in class can be broken down into only a few geons, but in order for the theory to be relevant, the world would have to break down into a reasonable number of geons. This is where the theory breaks down for me. Considering the complexity of the objects in the world, I cannot imagine that anything less than several million geons would be able to cover every object that we are able to identify, and with so many, the point of the theory, that the world is easily classifiable, is gone.
The Gestalt theory makes more sense to me, but even it makes the process of object recognition seem too simple. The Gestalt theory says that objects are grouped by proximity, similarity, closure, good continuation, and common fate. This makes sense to me because when we see a complex scene, we are able to see most things as different objects, or belonging to a group, like different people walking in a crowd, or a flock of birds flying through the sky. I think that especially with regards to good continuation and common fate the Gestalt principles make some very good points about our object recognition systems, however, even though I think the theory makes some good points, and makes much more sense than the geon theory, I still think that the gestalt theory is too simple. I have a hard time going from identifying a general object in a certain area to being able to identify your cousin in a picture of his third grade class.
All in all I think that both of the theories have pros and cons associated with them. While I do not have a better suggestion than either of these for how we are able to recognize and identify objects, I think that other theories, perhaps incorporating aspects of both, would give a better general description of how our visual system works.
March 30, 2008
