Image compression using iterated function systems and revolutionary programming: image compression without image metrics

摘要

A method which generates a compressed form for an input image without explicit reference to any complete image metrics is presented. The compressed form of the image is an iterated function system (IFS). The IFS is generated using an evolutionary programming optimization. This optimization simulates the competitive dynamics on an array of mapping cells. The result of the competition is a minimal number of parameters (mappings) which characterize the image. Because these parameters are calculated without using the overall image, it is believed that this method could lead to procedures for identifying objects without first segmenting them from their surroundings.