The algorithms of natural vision: the multi-channel gradient model

Nature, through the process of evolution, has developed strategies for the processing of visual information. These algorithms have been optimised over eons to maximise the organism's ability to survive in the real world. Hence, the techniques employed are efficient, environmentally robust and practical for implementation in the parallel architecture of the brain. Primate and human visual perception are the most developed, with a high percentage of cortical tissue devoted to interpreting the visual signal. Thus, we may examine the extensive neurophysiological and psychophysical evidence available in an attempt to decipher the algorithms used by biology in an effort to build artificial vision systems which incorporate many of the desirable traits of natural vision.