Cuckoo inspired fast search algorithm for fractal image encoding

Abstract

The search time and significant loss in compression are the significant constraints of the traditional fractal image compression. Hence the contemporary research contributions are aimed to discover optimal solutions to speed up the search speed with minimal loss of image significance at compression. Majority of the existing contributions achieve the search speed at the cost of decoded image quality and vice versa. In regard to this, we proposed a cuckoo inspired fast search (CIFS) technique for fractal image compression. Unlike the many of traditional models, which depend on 3 level wavelet classification, this proposed CIFS is using ordered vector of range blocks by their similarity and ordered vector of range blocks by their coordinate distance. The experimental study evinced that the proposed model is scalable and robust compared to PSO and GA based models found in contemporary literature. The significant reduction in mean square error calculations is also observed, since the only four transformations of the dihedral group are sufficient to compare for similarity here in this proposed CIFS. Keywords: Fractal, PSNR, Cuckoo search, PSO, Genetic algorithm, MSE