A Novel Algorithm for Bi-Level Image Coding and Lossless Compression based on Virtual Ant Colonies

Abstract

Ant colonies emerged as a topic of research and they are applied in different fields. In this paper, we develop an algorithm based on the concept of ant colonies and we utilize it for image coding and compression. To apply the algorithm on images, we represent each image as a virtual world which contains food and routes for ants to walk and search for it. Ants in the algorithm have certain type of movements depending on when and where they find food. When an ant finds food, it releases a pheromone, which allows other ants to follow the source of food. This increases the likelihood that food areas are covered. The chemical evaporates after a certain amount of time, which in turn helps ants move to cover another food area. In addition to the pheromone, ants use proximity awareness to detect other ants in the surrounding, which can help ants cover more food areas. When an ant finds food, it moves to that location and the movement and coordinates are recorded. If there is no food, an ant moves randomly to a location in the neighborhood and starts searching. We ran our algorithm on a set of 8 images and the empirical results showed that we could outperform many techniques in image compression including JBIG2.