The Problem

In many fields of human activity, machines have relieved humans of tasks that are an insult to human worth.  We have backhoes in place of ditch diggers, and computers in place of paper shufflers.  An enormous surge of creativity and caring has resulted from this liberation of human minds.

However, many human lives continue to be wasted in tasks that require little creativity or caring, but that require the enormous power of human vision.  There are two basic reasons why the power of human vision is necessary to these tasks:

  1. Humans, and the machines that work with them, must function in an environment that humans have created, and that is strongly influenced by human vision. 

  2. Human vision has evolved in a long and highly competitive process.  It is therefore the best available model of vision.

Human vision is harder for machines to take over than are other human faculties, because of its hidden complexity.  A person digs ditches with only two arms, but uses around a billion neurons, all functioning at once, for vision.  This leads to two basic problems for machine vision:

  1. A set of a billion processors is a steep requirement for any machine that is to be manufactured in significant quantities. 

  2. Getting that many processors to work together coherently is even more difficult.

Forward to Field Theory in Vision.
Up to How Quadrupole Convolution Works.