Genetic Algorithm-Based Image Approximation Using Triangle Representation for Efficient Compression

This paper presents a novel image approximation algorithm using genetic algorithms (GA) to achieve extreme compression for black-and-white images. The method optimizes the shape and placement of a limited number of triangles to approximate the target image, achieving recognizable results with as few as 139 bytes of data. The genetic algorithm’s evolutionary process is leveraged to iteratively improve image quality through mutation, crossover, and selection operators. The approach offers a highly compact representation of images, demonstrating significant size reductions compared to traditional image compression algorithms. The proposed technique balances visual fidelity with extreme data efficiency, making it a promising method for lossy image compression in constrained environments. Test results on sample images demonstrate the method’s effectiveness and potential applications.