A New Chain Code for Bi-Level Image Compression Using an Agent-Based Model of Echolocation in Dolphins

The tremendous increase in the size of data generated by the Internet of Things and other platforms has created an urgent need to investigate new methods for data representation and compression. In this article, we develop a novel system that is inspired by the echolocation system in dolphins for bi-level image encoding and employ it in data compression. To this end, we consider building an agent-based model that imitates the communications between different agents. Our agent-based modeling environment corresponds to an actual image, and it consists of dolphins and fish. The dolphins use echolocation to help them identify fish within certain distances around them. Additionally, the dolphins work within fixed ranges while encoding the needed information. The algorithm stores the initial coordinates of the dolphins' locations along with their movements and uses this information in image reconstruction. The experimental findings showed that we could significantly outperform well-known benchmarks in bi-level compression using the sample images we employed for testing.